# Local vs cloud meeting AI
Source: https://docs.stenoai.co/compare/local-vs-cloud
A comparison of local and cloud-based AI meeting recorders: privacy, accuracy, cost, and the trade-offs of each approach.
AI meeting tools fall into two categories: those that process audio in the cloud and those that process it on your device. The right choice depends on your privacy requirements, hardware, and how much you trust third parties with your recordings.
## How each approach works
**Cloud-based meeting AI** (Otter.ai, Fireflies.ai, Fathom, Grain, etc.)
Your audio or a live audio stream is sent to the vendor's servers. Transcription and summarization happen remotely and the results are returned to you. The vendor processes, stores, and may use your audio data under the terms of their privacy policy.
**Local meeting AI** (Steno)
By default, all processing happens on your Mac using models downloaded during setup. No audio leaves your device. Transcription uses two on-device engines -- Parakeet (the default, with a live transcript while you record) and Whisper (optional, chosen in Settings). Summarization uses a local language model running via [Ollama](https://ollama.com) by default, with an optional cloud model (OpenAI, Anthropic, or AWS Bedrock -- each billed by that provider) or a custom endpoint (which may be a paid API or your own self-hosted server) as an alternative -- in which case your transcript is sent to it. In the default local configuration there are no ongoing costs and no vendor relationship for your audio data.
## Side-by-side comparison
| | Steno (local) | Cloud meeting AI |
| ----------------------------------- | ---------------------------------------------- | ---------------------------- |
| **Audio stays on device** | Yes -- never transmitted | No -- sent to vendor servers |
| **Works offline** | Yes | No |
| **Account required** | No | Yes |
| **Cost** | Free for personal & team (open source) | $10-$30/month typically |
| **Transcription accuracy** | High (Parakeet / Whisper) | High (varies by vendor) |
| **Supported languages** | European on the default engine; 99 via Whisper | Varies (usually 30-60) |
| **Speaker labels** | Yes (`[You]` / `[Others]`) | Yes (varies) |
| **Meeting bot joins your call** | No | Often yes |
| **Other participants notified** | No | Sometimes |
| **Works with any call platform** | Yes | Varies |
| **Works in-person** | Yes | Depends on product |
| **Internet required** | No (after setup) | Yes, always |
| **Data used to train models** | No | Check vendor policy |
| **Suitable for confidential audio** | Yes | Requires careful review |
## When to use a cloud recorder
Cloud tools have real advantages in certain situations:
* You need real-time transcription shared with other meeting participants
* You want automatic calendar integration and bot joining
* Your team needs a shared workspace for meeting notes
* Accuracy for a specific language or accent is critical and you need to evaluate vendors
## When to use Steno
* Your recordings contain confidential, regulated, or privileged information
* You do not want a meeting bot appearing in your calls
* You want zero ongoing cost
* You work offline or in environments with restricted internet access
* You prefer to keep your data entirely under your own control
## Accuracy
Steno transcribes on-device with two engines: Parakeet (the default) and Whisper (optional, selected in **Settings → Transcribe**). Accuracy is high across common recording conditions. Parakeet covers 25 European languages, including English, and produces a live transcript while you record; switch to Whisper when you need one of its 99 supported languages.
***
Yes. Steno is independent of your call platform. You can run Steno for local notes while also using a team-facing cloud tool for shared notes -- they do not conflict.
On Apple Silicon, transcription is fast and runs entirely on-device, and the default Parakeet engine shows a live transcript while you record. Cloud tools also return transcripts in roughly real-time. Local processing is fast enough for most workflows, and everything stays on your Mac.
Cloud recording vendors have detailed privacy policies, but they generally involve storing your audio on their servers, using it for service improvement, and sharing it with subprocessors. Review the specific policy for any cloud tool you use with confidential audio.
# Steno vs Fathom
Source: https://docs.stenoai.co/compare/steno-vs-fathom
A direct comparison of Steno and Fathom: on-device vs cloud processing, cost, privacy, and which meeting notetaker to choose on a Mac.
Fathom is a cloud-based AI meeting notetaker with a generous free tier. Steno is a free, open-source macOS app that runs entirely on your device. Both record, transcribe, and summarize meetings — Fathom in the cloud, Steno on your Mac.
## At a glance
| | Steno | Fathom |
| -------------------- | ----------------------------------------- | -------------------------------------------------------- |
| **Audio processing** | On your Mac, always | Fathom's cloud (AWS, US) |
| **Transcription** | On-device (Parakeet / Whisper) | Cloud |
| **Summarization** | On-device (local model) or optional cloud | Cloud (OpenAI, Anthropic) |
| **Account required** | No | Yes |
| **Works offline** | Yes | No |
| **Meeting bot** | No | Optional (bot or bot-free capture) |
| **Price** | Free (personal & team) | Free tier / $20/mo Premium / $19-\$34 per user/mo (team) |
| **Open source** | Yes (MIT) | No |
| **CRM sync** | No | Yes (Salesforce, HubSpot) |
| **Platform** | macOS (stable), Windows 10/11 (alpha) | macOS, Windows, web |
*Fathom details as of 2026; check fathom.ai for current pricing and features.*
## The core difference: on-device vs cloud
Fathom offers both a traditional meeting bot and a newer bot-free desktop capture. Either way, it is a cloud service: your meeting audio is uploaded to Fathom's servers and processed with cloud AI providers, and your recordings live in your Fathom account.
Steno keeps everything on your Mac. Audio is transcribed and summarized locally, and no meeting content is uploaded unless you opt into a cloud summarization model. (Anonymous usage analytics -- no meeting content -- are sent by default and can be turned off in Settings.) If your requirement is that meeting content never leaves your device, that is the deciding factor.
## Cost
Fathom's free tier is genuinely useful — unlimited recording and transcription — but its AI value (advanced summaries, action items, the meeting assistant) sits behind Premium ($20/month) and team plans ($19-\$34/user/month) as of 2026. Steno is free for personal and team use and open source across all features; the only cost is disk space for the local models (roughly 7GB for the default setup).
## Where Fathom is stronger
* **CRM and sales workflows** — Fathom syncs to Salesforce and HubSpot and offers deal views, coaching metrics, and scorecards.
* **Cross-platform and team features** — shared search, playlists, comments, SSO, and a web app.
* **Zero setup for cloud users** — recordings and summaries are delivered to your inbox and searchable in the cloud.
## When to choose Steno
* You want transcription and summarization to happen entirely on your Mac
* You prefer no account and fully offline operation
* You want free, open-source software with no per-seat cost
* Your meetings are confidential and cannot be uploaded to a cloud service
## When to choose Fathom
* You need CRM sync and sales-coaching features
* Your team needs shared, searchable cloud recordings
* You work across Mac, Windows, and the web
* Cloud processing of meeting content is acceptable for your work
***
No. Fathom's bot-free capture removes the visible bot from the call, but the audio is still uploaded to Fathom's cloud for transcription and summarization. Steno processes everything on your Mac.
Steno is free for personal and team use and open source with no per-seat cost. Fathom has a usable free tier, but its AI features require a paid plan. Both avoid a meeting bot if you want one.
# Steno vs Fireflies.ai
Source: https://docs.stenoai.co/compare/steno-vs-fireflies
A direct comparison of Steno and Fireflies.ai: privacy, cost, features, and which is the right fit for your workflow.
Fireflies.ai is a cloud-based meeting recording and intelligence platform. Steno is a local macOS app. Both transcribe and summarize meetings, but they differ significantly in how they handle your audio and who they are designed for.
## At a glance
| | Steno | Fireflies.ai |
| ----------------------------------- | ---------------------------------------------- | ----------------------------------- |
| **Audio processing** | On your Mac | Fireflies' servers |
| **Account required** | No | Yes |
| **Price** | Free (personal & team) | Free (limited) / $10-$19/user/month |
| **Works offline** | Yes | No |
| **Meeting bot** | No | Yes (Fred bot) |
| **Bot visible to participants** | No | Yes |
| **Supported platforms** | macOS (stable), Windows 10/11 (alpha) | Web, Zoom, Meet, Teams, etc. |
| **Transcription** | Parakeet / Whisper (local) | Proprietary cloud |
| **Languages** | European on the default engine; 99 via Whisper | \~60 |
| **Speaker labels** | Yes | Yes |
| **CRM integration** | No | Yes (Salesforce, HubSpot, etc.) |
| **Search across all meetings** | On-device | Cloud |
| **Suitable for confidential audio** | Yes | Requires review |
## Privacy
Fireflies records meetings via its "Fred" bot, which joins your calls and records audio server-side. The audio is processed on Fireflies' infrastructure and stored in your Fireflies account.
Steno records audio on your Mac and processes it entirely locally in its default configuration. Nothing is transmitted to any server unless you explicitly configure a cloud summarization model (off by default). If your meetings contain information you cannot send to a third party, use Steno with the default local model.
## Meeting bot vs local recording
Fireflies' bot joins your video call as a visible participant. Attendees can see and hear that Fred has joined. This is standard for enterprise AI notetakers, but it changes the dynamic of the meeting and requires attendee awareness.
Steno captures audio from your Mac's audio system silently. No bot appears. The other participants are not notified by the app -- consent obligations remain yours to manage.
## CRM and integrations
Fireflies has deep integration with sales and business workflows: Salesforce, HubSpot, Slack, and others. It is widely used in sales and customer success teams where meeting notes need to flow into a CRM automatically.
Steno has no CRM integration. Its output is plain Markdown files, which you can route into other tools manually or via automation.
## Cost
Fireflies' free plan limits monthly transcription credits. The Pro plan is $10/user/month and Business is $19/user/month.
Steno is free for personal and team use and open source.
## When to choose Steno
* Your meetings contain confidential or privileged information
* You do not want a visible bot in your calls
* You want free, unlimited local transcription
* You are a solo user on macOS and do not need team features
## When to choose Fireflies.ai
* You need automatic CRM sync for sales or customer calls
* Your team needs a shared meeting intelligence platform
* You rely on integrations with tools like Slack, Notion, or HubSpot
* You need cross-platform recording (Zoom, Google Meet, Teams, etc.)
***
Fireflies stores recordings and transcripts in your account. Retention policies depend on your plan. Review Fireflies' privacy policy for current details on data storage and deletion.
Yes. Steno captures system audio from your Mac, which includes the audio output from any call platform -- Zoom, Google Meet, Teams, or any other app playing audio through your Mac's speakers. The other participants do not see any bot.
# Steno vs Granola
Source: https://docs.stenoai.co/compare/steno-vs-granola
A direct comparison of Steno and Granola: on-device vs cloud processing, cost, privacy, and which AI meeting notetaker to choose on a Mac.
Granola is a popular AI meeting notepad for Mac. Like Steno, it captures your computer's audio without sending a bot into the call. The key difference is where the work happens: Steno transcribes and summarizes **on your Mac**, while Granola relies on the cloud to generate your AI notes and stores them in your Granola account.
## At a glance
| | Steno | Granola |
| ----------------------------- | ----------------------------------------- | --------------------------------------------------- |
| **Audio processing** | On your Mac, always | Captured locally, processed in the cloud |
| **Transcription** | On-device (Parakeet / Whisper) | On-device or cloud |
| **AI notes / summaries** | On-device (local model) or optional cloud | Cloud (OpenAI, Anthropic) |
| **Notes stored** | On your Mac | Granola's cloud (US) |
| **Account required** | No | Yes |
| **Works offline** | Yes | No |
| **Meeting bot** | No | No |
| **Price** | Free (personal & team) | Free tier (limited history) / $14 / $35 per user/mo |
| **Open source** | Yes (MIT) | No |
| **Platform** | macOS (stable), Windows 10/11 (alpha) | macOS, Windows, iOS, Android |
| **Data used for AI training** | No | Yes, unless you opt out |
*Granola details as of 2026; check granola.ai for current pricing and policies.*
## The core difference: bot-free is not the same as local
Granola markets itself as bot-free, and it is — no participant joins your call. But bot-free describes only how audio is captured, not where it is processed. Granola relies on cloud AI to generate your notes, and your transcripts are stored in Granola's cloud.
Steno is bot-free **and** local: transcription and summarization run on your Mac, and no meeting content is uploaded unless you deliberately choose an optional cloud summarization model. (Steno does send anonymous usage analytics by default -- no meeting content -- which you can turn off in Settings.) For meetings where the audio itself cannot leave your device, that distinction matters.
## Cost
Granola offers a free tier with limited note history, then Business ($14/user/month) and Enterprise ($35/user/month) as of 2026. Steno is free for personal and team use and open source; the only cost is disk space for the local models (roughly 7GB for the default setup).
## Where Granola is stronger
* **Cross-platform and mobile** — Granola runs on Windows, iOS, and Android as well as Mac; Steno is macOS (stable) with a Windows 10/11 (alpha) build.
* **Team and CRM workflows** — Granola offers shared folders and paid integrations (Notion, Slack, HubSpot, and others).
* **Cloud sync** — your notes are available across devices automatically.
## When to choose Steno
* You want transcription and summarization to happen entirely on your Mac
* You prefer no account and fully offline operation
* You want free, open-source software you can inspect
* Your meetings are confidential and the audio cannot go to a cloud service
## When to choose Granola
* You need your notes on Windows or mobile as well as Mac
* You want cloud sync across devices and team sharing
* Cloud processing of meeting audio is acceptable for your work
***
No. Granola captures audio without a bot, but it relies on cloud AI to generate your notes and stores them on Granola's servers. Steno runs the whole pipeline on your Mac by default.
Yes. Steno has no account and works fully offline after the one-time model download. Granola requires signing in and stores your notes in its cloud.
# Steno vs MacWhisper
Source: https://docs.stenoai.co/compare/steno-vs-macwhisper
A direct comparison of Steno and MacWhisper: two local, on-device Mac transcription tools — cost, open source, and meeting features.
MacWhisper and Steno are the two closest tools in this comparison: both run on-device on a Mac, both avoid meeting bots, and both are built on open-source speech models. The differences are cost and licensing, and scope — MacWhisper is a paid, closed-source transcription app that has added meeting features, while Steno is free, open source, and purpose-built as a meeting recorder and summarizer.
## At a glance
| | Steno | MacWhisper |
| -------------------- | ----------------------------------------- | -------------------------------------------------------- |
| **Audio processing** | On your Mac | On your Mac |
| **Transcription** | On-device (Parakeet / Whisper) | On-device (Whisper, Parakeet) |
| **Summarization** | On-device (local model) or optional cloud | Optional (local or cloud AI add-ons) |
| **Meeting bot** | No | No |
| **Price** | Free (personal & team) | Free tier / one-time Pro license (\~€64) or subscription |
| **Open source** | Yes (MIT) | No (built on open-source Whisper) |
| **Focus** | Meeting recorder + summarizer | General transcription (files, meetings, media) |
| **Works offline** | Yes | Yes (core transcription) |
| **Platform** | macOS (stable), Windows 10/11 (alpha) | macOS only |
*MacWhisper details as of 2026; check macwhisper.com for current pricing and features.*
## Both are local and private
This is the one comparison where privacy posture is similar. MacWhisper transcribes on-device using Whisper and Parakeet, records system audio, and does not put a bot in your call — much like Steno. Both keep your audio on your Mac for the core transcription work, and both offer optional cloud AI features you can choose to turn on.
## The differences: cost, licensing, and scope
* **Cost and licensing** — Steno is free for personal and team use and open source (MIT), so you can inspect exactly how it handles your data. MacWhisper is a paid, closed-source app: a free tier plus a one-time Pro license (about €64 as of 2026) or a subscription via the Mac App Store. It is built on the open-source Whisper model, but the app itself is not open source.
* **Scope** — MacWhisper is a broad transcription tool: batch-transcribe files, transcribe media URLs, export subtitles, and more. Steno is purpose-built for meetings: record a call, get a live transcript, and get structured notes and action items from on-device summarization, with editable report templates.
## When to choose Steno
* You want a free, open-source tool you can audit
* You want on-device meeting summaries and notes, not just a transcript
* You want a purpose-built meeting recorder (live transcript, templates, ask-your-meetings)
## When to choose MacWhisper
* You transcribe a wide range of files and media, not only meetings
* You want batch transcription, subtitle export, or media-URL transcription
* A paid, closed-source app with a one-time license suits you
***
Yes — both run their core transcription on your Mac and avoid a meeting bot. The main differences are that Steno is free for personal and team use and open source with on-device meeting summarization built in, while MacWhisper is a paid, closed-source transcription tool with a broader file/media focus.
Steno is purpose-built for meetings: it records the call, shows a live transcript, and generates structured summaries and action items on-device, with editable templates. MacWhisper can transcribe meetings and add AI summaries, but its heritage is general-purpose transcription.
# Steno vs Otter.ai
Source: https://docs.stenoai.co/compare/steno-vs-otter
A direct comparison of Steno and Otter.ai: privacy, cost, transcription accuracy, and which to choose for your needs.
Otter.ai is a cloud-based AI meeting transcription service. Steno is a local macOS app. They solve the same core problem -- capturing and summarizing spoken meetings -- but in fundamentally different ways.
## At a glance
| | Steno | Otter.ai |
| ----------------------------------- | ------------------------------------------------------------------------------- | -------------------------------------- |
| **Audio processing** | On your Mac, always | Otter's servers |
| **Account required** | No | Yes |
| **Price** | Free (personal & team) | Free tier (limited) / $16.99-$30/month |
| **Works offline** | Yes | No |
| **Meeting bot** | No | Yes (OtterPilot) |
| **Other participants see the bot** | No | Yes |
| **Transcription engine** | Parakeet / Whisper (on-device) | Proprietary cloud |
| **Supported languages** | 25 European languages, including English, on the default engine; 99 via Whisper | Primarily English |
| **Speaker labels** | Yes | Yes |
| **Shared team workspace** | No | Yes |
| **Mobile app** | No (desktop only) | Yes (iOS, Android) |
| **Calendar integration** | Native (Google & Outlook auto-detect), plus Apple Shortcuts | Yes (native) |
| **Data used for training** | No | Check current policy |
| **Suitable for confidential audio** | Yes | Requires review |
| **Platform** | macOS (stable), Windows 10/11 (alpha) | Cross-platform |
## Privacy
Otter.ai processes audio on its servers. When you use Otter, your recording is transmitted to Otter's infrastructure, transcribed there, and stored in your Otter account. Otter's privacy policy governs what happens to that data.
Steno processes everything on your Mac. No audio is transmitted anywhere. This makes Steno suitable for meetings where sending audio to a third party is not acceptable -- clinical consultations, legal calls, financial discussions.
## Cost
Otter.ai's free tier limits monthly transcription minutes and restricts features. The Pro plan costs $16.99/month and Business is $30/user/month.
Steno is free for personal and team use and open source. The only cost is disk space for model files (roughly 7GB for the default setup).
## Meeting bot
Otter's OtterPilot feature joins your video calls as a visible participant. Other attendees will see "Otter" or "OtterPilot" in the participant list and receive a notification that the meeting is being recorded by an AI notetaker.
Steno has no meeting bot. It records audio from your Mac's audio system -- the other participants never see it in the call.
## Accuracy and language support
Both tools offer high transcription accuracy for clear, English-language audio. Steno transcribes on-device with two engines -- Parakeet (the default) and Whisper (optional) -- and performs well across accents and in noisy environments.
Otter is primarily optimized for English. Steno's default Parakeet engine covers 25 European languages, including English; switching to the Whisper engine adds automatic detection across 99 languages.
## Team features
Otter.ai includes shared workspaces, team folders, and collaborative notes -- features Steno does not have. If your team needs a shared repository of meeting notes with access controls, Otter is more suited to that use case.
Steno stores notes as plain Markdown files on your device. You can sync them with any tool you use for file sharing.
## When to choose Steno
* You work with confidential audio that cannot go to a cloud service
* You do not want a bot appearing in your calls
* You want free, unlimited transcription
* You work primarily on a Mac
## When to choose Otter.ai
* Your team needs a shared meeting notes workspace
* You need mobile recording (iOS or Android)
* You need live captions or real-time transcripts shared during a meeting
* Your meetings are predominantly in English and cloud storage is acceptable
***
Yes. They do not conflict. Some users run Steno for sensitive or private meetings and use Otter for general team meetings where a shared transcript is useful.
Both use strong transcription engines and perform similarly for clear, English-language audio. Steno's main advantage is that it runs on-device without sending audio anywhere; its default Parakeet engine covers 25 European languages, including English, and the optional Whisper engine adds language breadth (99 languages vs Otter's primarily English focus).
# Steno vs tl;dv
Source: https://docs.stenoai.co/compare/steno-vs-tldv
A direct comparison of Steno and tl;dv: on-device vs cloud processing, meeting bots, cost, and which notetaker to choose on a Mac.
tl;dv is a cloud-based AI meeting recorder and sales-intelligence platform. Steno is a free, open-source macOS app that records, transcribes, and summarizes meetings entirely on your device. The main difference is on-device versus cloud — and whether a bot joins your call.
## At a glance
| | Steno | tl;dv |
| ---------------------------------- | ----------------------------------------- | --------------------------------------------------------------- |
| **Audio processing** | On your Mac, always | tl;dv's cloud |
| **Transcription** | On-device (Parakeet / Whisper) | Cloud |
| **Summarization** | On-device (local model) or optional cloud | Cloud |
| **Account required** | No | Yes |
| **Works offline** | Yes | No |
| **Meeting bot** | No | Yes by default (Chrome); bot-free desktop mode available |
| **Other participants see the bot** | No | Yes, in bot mode |
| **Price** | Free (personal & team) | Free tier (10 AI notes lifetime) / paid from \~\$18 per user/mo |
| **Open source** | Yes (MIT) | No |
| **Platform** | macOS (stable), Windows 10/11 (alpha) | Web, Chrome, desktop (Mac/Windows), iOS |
*tl;dv details as of 2026; confirm current pricing on tldv.io — paid-tier prices vary across sources.*
## The core difference: on-device vs cloud, and the bot
By default, tl;dv records via a Chrome extension that sends a **visible bot** into your call. It also offers a bot-free desktop app that records system audio locally — but in both modes the recording and transcript are uploaded to tl;dv's cloud workspace and processed there.
Steno adds no bot: it records, transcribes, and summarizes on your Mac. No meeting audio, transcript, or summary leaves your device unless you opt into a cloud summarization model. Anonymous usage analytics (no meeting content) are sent by default and can be turned off in Settings.
## Cost
tl;dv's free tier lets you record but caps AI notes at 10 lifetime; ongoing AI value requires a paid plan (from roughly \$18/user/month billed annually, as of 2026). Steno is free for personal and team use and open source with no per-seat cost — only disk space for the local models (roughly 7GB for the default setup).
## Where tl;dv is stronger
* **Sales intelligence** — AI coaching, playbooks, and multi-meeting analysis across many calls.
* **CRM and automation** — HubSpot, Salesforce, Slack, Notion, and thousands of app integrations.
* **Cross-platform and team** — web app, mobile, and shared cloud workspaces.
## When to choose Steno
* You want everything to run on your Mac, with no upload
* You do not want a bot appearing in your calls
* You want free, open-source software with no per-seat cost
* Your meetings are confidential
## When to choose tl;dv
* You need sales coaching, CRM sync, and cross-meeting analytics
* Your team needs shared cloud recordings and integrations
* You work across web, Windows, and mobile
* Cloud processing is acceptable for your work
***
By default (the Chrome extension), yes — a visible tl;dv bot joins the call. Its desktop app can record without a bot, but the recording is still uploaded to tl;dv's cloud. Steno never adds a bot and never uploads.
No. tl;dv is a cloud service; recordings and transcripts are processed and stored on its servers, even in bot-free desktop mode. Steno processes everything locally on your Mac.
# Steno vs Zoom AI Companion
Source: https://docs.stenoai.co/compare/steno-vs-zoom-ai-companion
A direct comparison of Steno and Zoom AI Companion: on-device vs cloud, meeting coverage, cost, and privacy on a Mac.
Zoom AI Companion is the AI assistant built into Zoom Workplace. Steno is a free, open-source macOS app that records, transcribes, and summarizes meetings on your device. Zoom AI Companion is convenient if your meetings are all in Zoom; Steno works with any audio on your Mac and keeps everything local.
## At a glance
| | Steno | Zoom AI Companion |
| ----------------------------- | ------------------------------------------ | ---------------------------------------------------------- |
| **Audio processing** | On your Mac, always | Zoom's cloud (US) |
| **Transcription / summaries** | On-device | Cloud |
| **Meeting coverage** | Any call or in-person (system audio + mic) | Zoom-hosted meetings (Meeting Summary) |
| **Account required** | No | Yes — a paid Zoom plan + admin enablement |
| **Works offline** | Yes | No |
| **Price** | Free (personal & team) | Included with paid Zoom plans (e.g. Pro \~\$15.99/user/mo) |
| **Open source** | Yes (MIT) | No |
| **Platform** | macOS (stable), Windows 10/11 (alpha) | Zoom Workplace (Mac, Windows, mobile, web) |
*Zoom details as of 2026; check zoom.com for current plans. The separate "My notes" feature claims broader coverage but differs from Meeting Summary.*
## The core difference: ecosystem and location
Zoom AI Companion's meeting summaries work for meetings **hosted in Zoom** by a licensed user on a paid plan, with the feature enabled by an account admin. Processing happens in Zoom's cloud.
Steno is not tied to any meeting platform. It records the audio your Mac plays — Zoom, Teams, Google Meet, a browser call, or an in-person conversation through the microphone — and transcribes and summarizes it on your device by default. There is no account, no admin, and no meeting content is uploaded unless you opt into a cloud summarization model. (Steno does send anonymous usage analytics by default -- no meeting content -- with an opt-out in Settings.)
## Cost
Zoom AI Companion is included at no extra charge with paid Zoom accounts, but it requires one — it is not available on the free Zoom Basic plan (Zoom Workplace Pro is about \$15.99/user/month as of 2026). Steno is free for personal and team use and open source and works regardless of your Zoom plan.
## Where Zoom AI Companion is stronger
* **Native to Zoom** — no separate app; summaries and in-meeting Q\&A are built into the Zoom client.
* **Broad language support** and features across Team Chat, Phone, Whiteboard, and Mail.
* **Cross-platform** — wherever Zoom Workplace runs.
## When to choose Steno
* Your meetings are not all in Zoom (Teams, Meet, in-person, or mixed)
* You want processing to stay on your Mac, with no cloud upload
* You do not have — or do not want to pay for — a qualifying Zoom plan
* You want free, open-source software
## When to choose Zoom AI Companion
* Essentially all your meetings are hosted in Zoom
* You already pay for a qualifying Zoom plan and your admin has enabled it
* Cloud processing within the Zoom ecosystem is acceptable
* You want AI features across the wider Zoom Workplace suite
***
Its Meeting Summary feature covers Zoom-hosted meetings. Zoom's separate "My notes" feature claims broader coverage, but it is a distinct feature. Steno works with any meeting or in-person audio on your Mac, regardless of platform.
It is included with paid Zoom plans but is not available on the free Basic plan and must be enabled by an admin. Steno is free for personal and team use and open source and does not depend on your Zoom subscription.
# FAQ
Source: https://docs.stenoai.co/faq
Answers to common questions about Steno -- installation, privacy, recordings, and models.
## About Steno
Yes. Steno is free and open source under the MIT license, which places no restriction on who can use it or at what scale -- including enterprise and commercial use. The source is available on GitHub, so you can read exactly what it does, and there is no account to create. On top of the free app, a paid managed Enterprise deployment tier (admin controls, organization-specific features) is a separate, optional offering — details coming soon.
Yes. After the first-run model download, Steno works fully offline -- transcription and summarization run on-device, so no meeting audio, transcript, or summary is ever sent to a server. The app periodically checks for updates and sends anonymous usage analytics (no meeting content) by default; both can run without affecting offline recording, transcription, and summarization.
Steno runs on macOS (Apple Silicon, M1 and later -- Intel Macs are not supported since v0.4.0) and on Windows 10/11 (x64) in **alpha**, with the full on-device pipeline verified working. There is no iPhone or web version.
Steno saves your notes as plain Markdown files under `~/Library/Application Support/stenoai/output/` (macOS) or `%APPDATA%\stenoai\output\` (Windows). Because they are ordinary files, you can move them into Notion, Obsidian, or any other tool by hand or through your own automation.
Everything Steno stores lives under `~/Library/Application Support/stenoai/` (macOS) or `%APPDATA%\stenoai\` (Windows) -- recordings, transcripts, and summaries. To back up your data, copy that folder. To delete it, remove that folder. Nothing is stored remotely, so there is no cloud copy to clear.
## Privacy and data
No. Steno processes all recordings using AI models that run locally on your Mac. No audio, transcript, or summary is transmitted to any external server unless you configure an off-device summarization backend (cloud model, private server, or organization adapter). The app also sends anonymous usage analytics by default -- see the telemetry question below.
All data is stored locally:
**macOS**: `~/Library/Application Support/stenoai/`
**Windows**: `%APPDATA%\stenoai\` (e.g. `C:\Users\\AppData\Roaming\stenoai\`)
* `recordings/` -- raw `.webm` audio files
* `transcripts/` -- verbatim transcript text
* `output/` -- Markdown summaries and notes
Nothing is stored remotely.
Yes, anonymous usage analytics are on by default -- event names (e.g. "recording started," "summarization completed"), error types, and an anonymous install ID. This never includes audio, transcripts, notes, or summaries. Turn it off in **Settings → Advanced**.
Steno is designed for professionals handling confidential audio. Because processing is entirely local, it is suitable for use cases where sending audio to a cloud service is not acceptable -- clinical consultations, legal calls, financial meetings. Review your organization's specific policies before using any recording tool with regulated data. See [Confidential use cases](/privacy/confidential-use-cases) for more detail.
## Installation and setup
Steno is signed but distributed outside the Mac App Store. On first launch, macOS Gatekeeper may block it. Go to **System Settings → Privacy & Security** and click **Open Anyway** next to Steno. This is a one-time step.
Steno's Windows alpha builds are not digitally signed yet. Windows SmartScreen flags unsigned executables with a warning. To proceed, click **More info** and then **Run anyway**. The app is safe to run — we will sign the installer before the general availability release.
The default setup uses roughly 7GB: the Parakeet transcription model (\~572MB) plus the default Gemma summarization model. Plan for around 8GB free during install. Larger summarization models take more space. Recordings accumulate over time in `~/Library/Application Support/stenoai/recordings/` (macOS) or `%APPDATA%\stenoai\recordings\` (Windows).
Steno updates automatically in the background. When an update is ready, you will see a notice in the app. The update installs on the next quit -- you do not need to download a new DMG.
Delete Steno from your Applications folder (macOS) or uninstall via **Settings → Apps** (Windows). To also remove all recordings, transcripts, and model files, delete `~/Library/Application Support/stenoai/` (macOS) or `%APPDATA%\stenoai\` (Windows), and `~/.ollama/` (if you want to remove the Ollama models too).
## Recording
Yes. Steno captures your Mac's system audio and your microphone, so it is independent of the calling app -- Zoom, Microsoft Teams, Google Meet, or anything else playing audio on your Mac. In-person meetings work too, captured through the microphone.
No. Steno records audio locally from your Mac's audio system and never joins the call as a participant. No bot appears, and no other attendee is notified by Steno. Recording-consent obligations remain your responsibility.
Steno can capture system audio -- the audio playing through your Mac's speakers or headphones, which includes the other participants on a call. See [Recording](/features/recording) for setup instructions. System audio capture requires a one-time permission setup.
Yes. Steno records from your Mac's microphone, which will pick up voices in the room. For best results, place your laptop centrally. In-person recordings are mic-only, so they have no `[You]` / `[Others]` speaker labels.
No. Steno captures audio from your Mac's audio system without joining the call as a participant. Other attendees will not see a bot or receive a notification from Steno. Recording consent obligations are your responsibility.
## Transcription and summaries
Steno defaults to the Parakeet engine, which is fast on Apple Silicon and shows a live transcript while you record. If you have switched to the Whisper engine in **Settings → Transcribe**, switching back to Parakeet is usually faster; Whisper transcribes only after you stop recording.
Switch to a larger summarization model. In **Settings → AI**, try `qwen3.5:9b`, `gemma4:12b-it-qat`, or `gpt-oss:20b`. Larger models produce more detailed and accurate notes. You can regenerate notes for any existing recording from its detail view.
Yes. The default Parakeet engine transcribes 25 European languages, including English, French, German, Spanish, Italian, Dutch, Portuguese, Russian, and Polish. A curated subset of six (English, French, German, Spanish, Dutch, Portuguese) is available to pin as the summary output language. For Japanese, Chinese, Korean, Hindi, Arabic, and the full set of 99 languages, switch to the Whisper engine in **Settings → Transcribe**, which auto-detects the language from the audio.
## Models
Transcription: two on-device engines -- Parakeet (the default) and [Whisper](https://github.com/ggerganov/whisper.cpp) (optional, its single `large-v3-turbo` model), selected in **Settings → Transcribe**. Summarization: a local model via [Ollama](https://ollama.com), with `gemma4:e2b-it-qat` as the default. See [Transcription models](/models/transcription-models) and [Summarization models](/models/summarization-models) for the full comparison.
Yes. Steno optionally supports OpenAI, Anthropic, AWS Bedrock, or a custom OpenAI-compatible endpoint as an alternative to the local model. Configure this in **Settings → AI**, under **Cloud API**. It is off by default. Note that this sends your transcript and any typed notes (not audio) to the selected provider's API.
# Apple Shortcuts
Source: https://docs.stenoai.co/features/apple-shortcuts
Automate meeting recordings on macOS using Apple Shortcuts and the Steno deep link protocol. Start and stop recordings from a keyboard shortcut, menu bar script, or calendar event.
Steno supports automation via the `stenoai://` deep link scheme. You can trigger recording start and stop from Apple Shortcuts, Automator, or any tool that can open URLs on macOS.
## Deep links
| Action | URL |
| --------------------------- | -------------------------------------------- |
| Start recording | `stenoai://record/start` |
| Start recording with a name | `stenoai://record/start?name=Meeting%20Name` |
| Stop recording | `stenoai://record/stop` |
The `name` parameter is URL-encoded. Spaces become `%20`.
## Create a start/stop shortcut
Open **Shortcuts** from your Applications folder or search for it with Spotlight.
Click the **+** button to create a new shortcut. Name it "Start Steno recording".
Search for "Open URLs" in the action search bar and add it to the shortcut.
Enter `stenoai://record/start` as the URL. To pass a name, use `stenoai://record/start?name=Daily%20Standup`.
In the shortcut's settings (top-right menu), assign a keyboard shortcut -- for example `⌃⌥R` -- so you can start a recording without opening any app.
Create a second shortcut for `stenoai://record/stop` and assign it a different key combination.
## Calendar automation
macOS Shortcuts cannot natively trigger at calendar event start times. To automate recordings based on your calendar, use [Rules - Calendar Automation](https://apps.apple.com/app/rules-calendar-automation/id1619042151) as a bridge.
In Shortcuts, create a shortcut that:
1. Uses **Find Calendar Events** (limit to 1, upcoming, sorted by start date)
2. Extracts the event title
3. URL-encodes the title
4. Opens `stenoai://record/start?name=[encoded title]`
In Rules, create a calendar trigger:
* **Source**: your target calendar
* **Trigger**: event start (or a few minutes before)
* **Condition**: event note contains `steno` (use this as an opt-in marker)
* **Action**: run your Steno shortcut
In Calendar, add the word `steno` to the notes field of any event you want automatically recorded. Events without this marker will not trigger Steno.
## Use with the menu bar
Steno runs a menu bar icon when active. You can trigger recordings from the menu bar without using Shortcuts. The deep links are most useful when you want to trigger recordings from other apps, scripts, or automations.
***
Yes. Steno must be running (in the menu bar is fine) to receive deep link events. If Steno is closed, the URL will launch it but may not immediately start recording on the first open.
Yes. URL-encode the name and pass it as the `name` query parameter: `stenoai://record/start?name=Weekly%20Product%20Sync`. Steno uses this as the recording title.
Yes. Both Raycast and Alfred can open URLs. Create a script command or workflow that opens `stenoai://record/start` or `stenoai://record/stop`.
macOS Shortcuts does not provide a trigger that fires at a calendar event's start time. Rules adds this capability. It watches your calendar and runs a Shortcut when an event starts.
# Ask your meetings
Source: https://docs.stenoai.co/features/ask-your-meetings
Query any saved meeting note in plain English. Steno answers from your local transcript and summary -- no data leaves your Mac.
Every meeting note in Steno has an inline ask bar. Type a question in plain English and Steno answers from the saved transcript, summary, and key topics using the local language model.
## How to use it
1. Open a saved note from the sidebar
2. Click the **Ask** bar at the bottom of the pane
3. Type your question and press Return
Steno reads the saved `.md` files for that meeting -- the summary, key topics, and full transcript -- and passes them with your question to the local model. The answer streams inline below the ask bar.
## What you can ask
The ask bar is useful for questions you would ask a colleague who was in the meeting:
* "What decisions were made?"
* "What did we agree to do about the pricing change?"
* "Who is responsible for the API migration?"
* "Summarize this in three bullet points."
* "What were the concerns raised about the timeline?"
* "Was there anything about \[client name]?"
Questions that reference specific details from the meeting work best. Vague questions ("what happened?") return less useful answers than specific ones ("what was the outcome of the budget discussion?").
## What Steno reads
When you ask a question, Steno reads three sources from the saved note:
| Source | Content |
| --------------- | ------------------------------------------------------------- |
| Summary | The AI-generated paragraph summary of the meeting |
| Key topics | The structured list of topics identified during summarization |
| Full transcript | The verbatim transcript with timestamps and speaker labels |
The full transcript is included, bounded by the active model's context window. Very long meetings can exceed that window; switching to a model with a larger context window (see [Summarization models](/models/summarization-models)) helps if a question about the end of a long meeting gets a weaker answer.
## Privacy
All queries are processed by the local model on your Mac. Your questions and the meeting content are not sent to any external server unless you have configured a cloud model in Settings.
***
Yes. Alongside the inline ask bar for a single note, Steno supports library-wide chat across all of your meetings, so you can ask questions that span your whole meeting history.
Larger models produce better answers. In **Settings → AI**, switch to `qwen3.5:9b` or `gemma4:12b-it-qat`. Also check that the full transcript was saved -- if transcription failed or was skipped, the model only has the summary to work from.
Yes. The ask bar passes your question directly to the language model, which supports multilingual input. Results may vary by model and language.
# Live transcription
Source: https://docs.stenoai.co/features/live-transcription
On macOS, Steno shows a live transcript while you record — text appears as people speak, entirely on-device. No audio leaves your Mac.
Live transcription shows the transcript building in real time while you record. Words appear on screen as people speak, processed entirely on your Mac. It runs on the **Parakeet** engine (the default on macOS), so there is nothing to enable — start a recording and the live transcript appears.
## How it works
While recording, Steno feeds audio through voice-activity detection and the Parakeet model to produce a running transcript. When you stop, the same recording is finalized into the saved transcript that your notes are built from. The live view is a convenience during the meeting; the saved transcript is the source of record.
When the model is still loading at the start of a recording, Steno shows that it is warming up while audio continues to be captured — nothing is lost.
## Engine requirements
Live transcription is available when the **Parakeet** engine is active (Settings → Transcribe). If you switch to the **Whisper** engine, recordings transcribe **after** you stop rather than live. See [Transcription models](/models/transcription-models) for the difference between the two engines.
***
When you stop recording, Steno finalizes the recording into the saved transcript under `~/Library/Application Support/stenoai/transcripts/`, and your notes are generated from it. The live view is what you see during the meeting.
No. Live transcription runs entirely on your Mac, the same as the rest of Steno's on-device pipeline.
# Recording
Source: https://docs.stenoai.co/features/recording
Record microphone audio, system audio, or both. Steno captures audio locally and never uploads your recordings.
Steno always records your microphone, with system audio (what you hear through your headphones or speakers) as an optional add-on you can toggle on or off.
## Microphone recording
Microphone recording is always on. It captures your voice and any voices in the same room as your Mac. No setup is required beyond granting microphone permission on first use.
Best for: in-person meetings, interviews, one-on-one conversations where you are the primary speaker.
## System audio recording
System audio capture records the audio playing through your Mac -- including the other participants on a Zoom, Google Meet, Teams, or any other call. Combined with microphone recording, this captures both sides of a virtual meeting.
**Setup (one time)**
macOS requires explicit permission for apps to record system audio. On first use:
1. Open the recording options popover (the **···** button next to the record button) and turn on **Record system audio**
2. If macOS prompts for permission, open **System Settings → Privacy & Security → Screen & System Audio Recording**
3. Enable Steno
4. Relaunch Steno if macOS asks you to
After this one-time setup, the toggle stays on for future recordings until you turn it off. With it off, Steno records the microphone only -- useful for in-person meetings or dictation.
## Speaker labels
When Steno captures system audio, it labels lines in the transcript as `[You]` (microphone audio) or `[Others]` (system audio). This makes it easier to follow who said what in a virtual meeting.
Speaker labels appear automatically -- no configuration is needed.
## Starting a recording
Click **New note** in the toolbar or use the menu bar icon -- recording starts immediately with an auto-generated title, which you can rename from the note's detail view at any time.
## Stopping a recording
Click **Stop** in the toolbar or from the menu bar icon. Transcription begins immediately.
## Apple Shortcuts
You can start and stop recordings remotely using Apple Shortcuts and the `stenoai://` deep link scheme. See [Apple Shortcuts](/features/apple-shortcuts) for the full setup guide.
***
Steno records in WebM/Opus format. By default, files are stored in `~/Library/Application Support/stenoai/recordings/`. You can change the storage location in **Settings → Advanced**.
Not currently. System audio capture records all audio playing on your Mac -- there is no per-app filtering.
No. Steno will record until you stop it or your disk fills up.
# Report templates
Source: https://docs.stenoai.co/features/report-templates
Steno generates meeting notes from editable prompt templates. Pick a built-in template, edit its prompt, or create your own — and keep more than one report per meeting.
A report template is a named, editable prompt that tells the summarization model how to write your notes. Steno ships a small gallery of built-in templates, lets you edit any of them, and lets you add your own — so the same recording can produce a sales-call recap, a 1:1 summary, or a standup update, each in its own format.
## Built-in templates
Steno includes a curated set of templates you can use as-is or edit:
| Template | Best for |
| ---------------- | ----------------------------------------------------------------- |
| **Standard** | General meetings — summary, key topics, key points, action items |
| **Product Demo** | Demo calls — what was shown, questions raised, follow-ups |
| **Sales Call** | Discovery and sales conversations — needs, objections, next steps |
| **1:1** | One-on-ones — discussion, decisions, action items |
| **Standup** | Team standups — updates, blockers, next steps |
## Editing a template
Every built-in template is editable. Open **Settings → Templates**, select a template, and edit its prompt. If you want the original back, **Reset** restores the shipped default. Your edits are stored locally and never leave your Mac.
You can also create a **custom template** from scratch — give it a name and a prompt, and it appears alongside the built-ins.
## Choosing the default
The **default template** is the one Steno uses automatically when a recording finishes and auto-summarize is on. Set it in **Settings → Templates**. New installs default to **Standard**.
## Multiple reports per meeting
A single meeting can hold more than one report. From a meeting's detail view you can generate an additional report with a different template, switch which report is shown as the active one, or delete a report you no longer need. Each report is saved alongside the meeting.
***
Templates live in your local config, and generated reports are saved alongside each meeting under `~/Library/Application Support/stenoai/`. No meeting content is uploaded unless you have configured a cloud summarization model.
Yes. A template can pin an output language, so notes are written in that language regardless of the meeting audio. Left on auto, the notes follow your configured summarization language.
# Speaker labels
Source: https://docs.stenoai.co/features/speaker-labels
Steno labels transcript lines as [You] or [Others] when system audio is enabled, making it easy to follow who said what in virtual meetings.
When Steno captures both microphone and system audio, transcripts include speaker labels: `[You]` for lines from your microphone and `[Others]` for audio from your speakers (the other participants).
## How it works
Steno records the microphone and system audio as two separate channels, then transcribes each channel and labels its lines -- the microphone channel `[You]`, the system audio channel `[Others]`. This stereo-channel diarization works the same way regardless of which transcription engine (Parakeet or Whisper) is active.
## Requirements
Speaker labels require **system audio capture** to be enabled -- microphone-only recordings produce a single-channel transcript without speaker differentiation.
See [Recording → System audio](/features/recording) for setup instructions.
## Limitations
* Labels are binary: `[You]` and `[Others]`. Individual speakers among "Others" are not distinguished.
* Label accuracy depends on audio separation. If microphone audio bleeds into system audio or vice versa, some lines may be mislabeled.
* In-person meetings (microphone only) do not support speaker labels.
Multi-speaker diarisation (identifying individual speakers by voice) is on the roadmap.
***
Not currently. Speaker labels distinguish only between your audio (`[You]`) and all other audio (`[Others]`). Individual speaker identification is planned for a future release.
# Summaries and notes
Source: https://docs.stenoai.co/features/summaries
After transcription, Steno generates structured meeting notes: a summary paragraph, key topics, key points, and action items -- from a local model by default, or a configured cloud model.
After each recording is transcribed, Steno generates structured notes using a language model. By default this is a local model running via [Ollama](https://ollama.com) on your device; you can optionally configure a cloud model instead (see [Summarization models](/models/summarization-models)). Notes are saved as Markdown files on your device either way.
## What Steno generates
Each note includes:
* **Summary** -- a short paragraph covering the main points of the meeting
* **Key topics** -- a bulleted list of the subjects discussed
* **Key points** -- the notable takeaways drawn from the discussion
* **Action items** -- tasks or commitments identified in the transcript
The quality and detail of notes depends on the model you are using. The default summarizer is Gemma 4 E2B; larger models produce more accurate and detailed output. See [Summarization models](/models/summarization-models).
## In-recording notes
You can type notes during a recording using the notes panel. These notes are folded into the AI summary when summarization runs -- Steno treats them as additional context alongside the transcript.
## Output format
Summaries are saved as Markdown files in `~/Library/Application Support/stenoai/output/`. They are plain text and work with any Markdown editor, note-taking app, or sync service.
## Regenerating a summary
You can regenerate the summary for any existing recording:
1. Open the recording
2. Click **Regenerate notes** in the toolbar
3. Optionally switch to a different model in **Settings → AI** before running
## Asking questions about a note
Use the [ask bar](/features/ask-your-meetings) at the bottom of any note to query the meeting content in plain English.
***
Yes. Report templates are editable directly in the UI. You can use the standard template or create custom ones, and generate multiple reports per meeting from different templates.
In `~/Library/Application Support/stenoai/output/` as `.md` files named after the recording.
# Transcription
Source: https://docs.stenoai.co/features/transcription
Steno transcribes recordings on-device with two engines: Parakeet (default) and Whisper (optional). Automatic language detection, live transcription, no audio uploads.
Steno transcribes with two on-device engines: **Parakeet** (the default, NVIDIA Parakeet TDT v3 running via MLX on Apple Silicon) and **Whisper** (an optional alternate, OpenAI's Whisper `large-v3-turbo`). Both run entirely on your device -- your audio is never sent to any server. You choose the engine in **Settings → Transcribe**.
## How it works
Transcription runs on-device and is fast on Apple Silicon. With the default Parakeet engine, a live transcript appears while you record, and Steno finalizes it when you stop. With the Whisper engine, transcription runs after you stop the recording. Either way, the active engine processes the audio file locally and outputs a verbatim transcript with timestamps.
## Language support
Language coverage depends on the engine. Steno auto-detects the language -- you do not need to set it manually.
* **Parakeet (default)** transcribes 25 European languages, including English, French, German, Spanish, Italian, Dutch, Portuguese, Russian, and Polish. A curated subset of six (English, French, German, Spanish, Dutch, Portuguese) can be pinned as the summary output language.
* **Whisper (optional)** covers the full set of 99 languages, including Japanese, Chinese, Korean, Hindi, and Arabic. To transcribe any of these, switch to the Whisper engine in **Settings → Transcribe**.
## Speaker labels
Transcripts include `[You]` and `[Others]` labels when system audio is enabled. `[You]` labels lines from your microphone; `[Others]` labels the audio from your speakers.
## Viewing the transcript
The full transcript appears in the right pane of any saved note, below the AI summary. Transcripts are also saved as plain text files in `~/Library/Application Support/stenoai/transcripts/`.
## Choosing an engine
See [Transcription models](/models/transcription-models) for a comparison of the Parakeet and Whisper engines.
***
Yes. On macOS with the default Parakeet engine, a live transcript appears while you record. If you switch to the Whisper engine, transcription runs after you stop the recording.
Transcripts are saved as plain text files. You can open and edit them in any text editor. Changes to the transcript file are reflected in Steno the next time you open the note.
# Your first recording
Source: https://docs.stenoai.co/getting-started/first-recording
Make your first recording with Steno. From pressing record to reading your summary takes about two minutes.
This guide walks through making a short test recording so you can verify the full pipeline is working: record → transcribe → summarize → query.
## Make a recording
Launch Steno from your Applications folder or from the menu bar icon.
Click **New note**. Steno starts capturing audio immediately and generates a title automatically from the transcript -- you can rename it later from the note's detail view.
You will see a live timer in the top bar, and with the default Parakeet engine a live transcript appears while you record.
Speak for 30-60 seconds. Introduce yourself, describe your work, or read a few sentences aloud -- anything with clear speech works.
Click **Stop**. Steno immediately begins transcribing.
## Review the output
Transcription runs on-device and is fast on Apple Silicon, so a short recording is ready in moments. When it finishes:
* The **transcript** appears in the right pane with timestamped lines
* A **summary** is generated below the transcript: a short paragraph, key topics, and any action items Steno identified
* The recording appears in the **sidebar** under today's date
## Ask a question
Click the **Ask** bar at the bottom of any note. Type a question in plain English:
* "What were the main points?"
* "What actions did I mention?"
* "Summarize this in one sentence."
Steno passes your question and the saved note to the local model and streams the answer inline. No data leaves your Mac when using the default local model.
## Where your files are saved
All output is stored locally:
| File type | Location |
| --------------- | ---------------------------------------------------- |
| Audio | `~/Library/Application Support/stenoai/recordings/` |
| Transcript | `~/Library/Application Support/stenoai/transcripts/` |
| Summary / notes | `~/Library/Application Support/stenoai/output/` |
Summaries and transcripts are plain Markdown files. You can open, edit, and sync them with any tool you use for notes.
***
Transcription speed depends on which engine you use. It runs entirely on-device and is fast on Apple Silicon; the default Parakeet engine also shows a live transcript as you record. The optional Whisper engine runs more slowly, since it transcribes only after you stop. You can switch engines in **Settings → Transcribe**.
Try a larger summarization model. In **Settings → AI**, switch from the default `gemma4:e2b-it-qat` to `qwen3.5:9b` or `gpt-oss:20b`. Larger models produce better-structured notes but take longer to run.
You need to enable system audio capture. See [Recording → System audio](/features/recording) for the one-time setup.
# Installation
Source: https://docs.stenoai.co/getting-started/installation
Download and install Steno on macOS. The setup takes about five minutes, most of which is downloading the local AI models.
## Requirements
* macOS 14.4 (Sonoma) or later, on Apple Silicon (M1 or newer) -- Intel Macs are not supported since v0.4.0; the last release with Intel support is [v0.3.8](https://github.com/ruzin/stenoai/releases/tag/v0.3.8)
* 8GB RAM minimum (16GB recommended for larger models)
* \~8GB free disk space for the default models
## Install Steno
Go to [stenoai.co](https://stenoai.co) and click **Download** to get the `arm64` build for Apple Silicon Macs (M1 through M5).
Open the downloaded `.dmg` file. Drag the Steno icon into your Applications folder.
Open Steno from Applications. macOS may show a security prompt on first launch -- click **Open** to proceed. This prompt appears because Steno is distributed outside the Mac App Store.
If macOS blocks the app entirely, go to **System Settings → Privacy & Security** and click **Open Anyway** next to the Steno entry.
On first launch, Steno will prompt you to download the default models:
* Transcription: Parakeet (\~572MB)
* Summarization: Gemma 4 E2B (\~6.5GB)
Both downloads happen inside the app. This requires an internet connection and takes a few minutes depending on your connection. After this, Steno runs entirely offline.
When you start your first recording, macOS will prompt for microphone permission. Click **Allow**. To record system audio (both sides of a call), follow the [system audio setup](/features/recording) guide.
## Verify the installation
Once setup is complete, go to **Settings → Setup Check** inside the app. This runs a diagnostic that confirms:
* The transcription model is installed and functional
* Ollama is running and the summarization model is loaded
* Storage paths are writable
If any check fails, the setup screen will show the specific error and how to resolve it.
## Updates
Steno updates automatically in the background. When an update is ready, you will see an **Update available** notice in the app. Updates install on the next quit -- you do not need to re-download a DMG manually.
***
You need an internet connection to download the initial model files. After that, Steno runs fully offline.
macOS Gatekeeper may block apps distributed outside the App Store. Go to **System Settings → Privacy & Security** and click **Open Anyway** next to Steno. This is a one-time step.
Delete Steno from your Applications folder. To also remove all data (recordings, transcripts, summaries, models), delete `~/Library/Application Support/stenoai/`.
# What is Steno?
Source: https://docs.stenoai.co/getting-started/what-is-steno
Steno is a macOS app that records, transcribes, and summarizes meetings entirely on-device. Your audio never leaves your Mac.
Steno is a macOS app that records, transcribes, and summarizes meetings using AI models that run entirely on your device. By default there are no cloud uploads and no accounts required; the only outbound requests are model downloads during setup, periodic update checks, and anonymous usage analytics (no meeting content) that you can turn off in Settings. Your audio, transcripts, and summaries are stored locally in `~/Library/Application Support/stenoai/` (you can change the location in Settings).
## How it works
Steno runs a local pipeline on every recording:
1. **Record** -- always captures microphone audio, with system audio (both sides of a call) as an optional add-on toggle
2. **Transcribe** -- runs a local speech-to-text engine to produce a verbatim transcript with speaker labels when system audio is on. Steno ships two engines: Parakeet (the default, which also shows a live transcript while you record) and Whisper (an optional alternate you can switch to in Settings)
3. **Summarize** -- passes the transcript to a small language model running via [Ollama](https://ollama.com) on your Mac to produce structured notes: summary, key topics, and action items
4. **Query** -- ask natural-language questions against any saved note from the inline ask bar
Recording, transcription, and summarization all run offline. Steno does not send meeting content to any external server unless you explicitly configure a non-local backend (a cloud model such as OpenAI, Anthropic, AWS Bedrock, or a custom OpenAI-compatible endpoint, or a remote Ollama / organization adapter) as an optional alternative to the local model.
## Who it's for
Steno is built for professionals who work with confidential audio:
* **Healthcare** -- clinical consultations, case reviews, MDT meetings
* **Legal** -- client calls, depositions, strategy sessions
* **Finance** -- investment committee meetings, client briefings
* **Any role** where recordings leaving your device is not acceptable
It is also used by developers, researchers, and individuals who simply prefer their data to stay on their own hardware.
## What Steno is not
Steno is not a meeting bot. It does not join your video calls as a participant, does not require calendar access, and does not require you to share a link with other attendees. It records audio from your Mac's audio system -- no bot appears and the app does not notify other participants. Consent obligations remain yours to manage under your local laws and policies.
## Supported platforms
macOS 14.4 (Sonoma) or later, on Apple Silicon (M1 and newer). Intel Macs are not supported since v0.4.0.
***
No, in the default local configuration. After the initial download and model installation, Steno runs offline -- transcription and summarization happen on your Mac. The app still performs periodic update checks and sends anonymous usage analytics (no meeting content, can be turned off in Settings), and if you opt into a cloud summarization model (OpenAI, Anthropic, AWS Bedrock, or a custom OpenAI-compatible endpoint) it will reach that provider's API for summary generation only.
All data is stored in `~/Library/Application Support/stenoai/`. This includes recordings (`recordings/`), transcripts (`transcripts/`), and summaries (`output/`). Nothing is stored remotely.
Yes. The source code is available at [github.com/ruzin/stenoai](https://github.com/ruzin/stenoai) under the MIT license.
An Apple Silicon Mac (M1 or later) with at least 8GB of RAM -- the default models total roughly 7GB on disk. Larger optional summarization models need more RAM and disk space.
# How to take notes in an in-person meeting on a Mac
Source: https://docs.stenoai.co/guides/in-person-meeting-notes
Record and summarize in-person meetings on macOS with Steno. Your Mac's microphone captures the room; transcription and notes are generated on-device.
To take notes in an in-person meeting on a Mac, use Steno with your microphone: it records the room through your Mac's mic, then transcribes on your device and summarizes locally by default. There is no meeting app and no bot involved — just the microphone — so it works for any face-to-face discussion.
## Steps
1. Install Steno and grant microphone permission (see [Installation](/getting-started/installation)).
2. In Steno, make sure the **Record system audio** toggle is off (microphone recording is always on), then click **New note** to start.
3. Place your Mac where its mic can hear the room clearly.
4. When the meeting ends, click **Stop**. Steno transcribes the recording and generates notes automatically.
## Speaker labels in person
Speaker labels (`[You]` / `[Others]`) come from recording your microphone and system audio as separate channels. An in-person meeting is captured on the microphone only, so it does not produce speaker labels — the transcript is a single stream. See [Speaker labels](/features/speaker-labels).
## Getting a clean transcript
Room audio can be noisier than a call. For in-person recordings where accuracy matters, the [Whisper engine](/models/transcription-models) can be a good choice. Position the Mac close to the speakers and minimize background noise where you can.
***
Not today. In-person audio is a single microphone stream, so it is transcribed as one channel. Multi-speaker diarisation (identifying individual speakers by voice) is on the roadmap.
No. In-person recordings are transcribed and summarized on your Mac. No meeting content is uploaded unless you have configured an optional cloud summarization model. The app does send anonymous usage analytics by default (no meeting content), which you can turn off in **Settings → Advanced**. See [How on-device processing works](/privacy/how-on-device-works).
# How to record a Google Meet call on a Mac
Source: https://docs.stenoai.co/guides/record-google-meet-on-mac
Record and transcribe Google Meet calls on macOS without a bot. Steno captures system audio and your microphone locally and writes the notes on-device.
To record a Google Meet call on a Mac, use Steno: it captures your Mac's system audio (the other participants) and your microphone together, then transcribes the call entirely on your device and summarizes it locally by default. No bot joins the meeting, and no meeting content is uploaded unless you configure a cloud summarization model. Google Meet runs in the browser, and Steno records the audio your Mac plays — so there is nothing extra to install in Meet.
## Steps
1. Install Steno and grant microphone and system-audio permission, and turn on the **Record system audio** toggle (see [Installation](/getting-started/installation) and [Recording](/features/recording)).
2. Join your Google Meet call in your browser as usual.
3. In Steno, click **New note** to start recording. It captures your microphone and system audio together.
4. When the call ends, click **Stop**. Steno transcribes the recording and generates notes automatically.
## Why there's no bot
Steno does not join your Google Meet call as a participant. It records the audio locally on your Mac, so no one sees a recording bot or gets a notification from Steno. Recording-consent obligations are your responsibility — see [Confidential use cases](/privacy/confidential-use-cases).
## Speaker labels
With system audio enabled, your microphone and the call audio are recorded as separate channels, so the transcript is labelled `[You]` and `[Others]`. See [Speaker labels](/features/speaker-labels).
***
No. Steno records your Mac's system audio directly, so it captures Google Meet without any browser extension or add-on.
No. Recording, transcription, and summarization all run on your Mac. No meeting content is uploaded unless you have configured an optional cloud summarization model. The app does send anonymous usage analytics by default (no meeting content), which you can turn off in **Settings → Advanced**. See [How on-device processing works](/privacy/how-on-device-works).
# How to record system audio on a Mac
Source: https://docs.stenoai.co/guides/record-system-audio-on-mac
Capture the audio your Mac plays — calls, videos, any app — with Steno. A one-time permission setup enables system-audio recording alongside your microphone.
To record system audio on a Mac, Steno uses a one-time audio-permission setup that lets it capture the sound your Mac plays back — the other side of a call, a video, or any app — at the same time as your microphone. Once enabled, every recording can include system audio, with your mic and the system audio kept as separate channels for speaker labels.
## One-time setup
1. In Steno, open the recording options popover (the **···** button next to the record button) and turn on **Record system audio**.
2. If macOS prompts for permission, open **System Settings → Privacy & Security → Screen & System Audio Recording**.
3. Enable Steno, then relaunch the app if macOS asks you to.
4. The toggle stays on for future recordings until you turn it off.
See [Recording](/features/recording) for the full recording options.
## What system audio captures
System audio is everything your Mac plays through its output — the remote participants on a call, a video you are watching, or audio from any app. Combined with your microphone, it lets Steno record both sides of a conversation. There is no per-app filtering; Steno captures the system output as a whole. Turn the toggle off to record the microphone only.
## Speaker labels
Because your microphone and the system audio are recorded as separate channels, the transcript is labelled `[You]` (your mic) and `[Others]` (system audio). See [Speaker labels](/features/speaker-labels).
***
macOS requires explicit permission before an app can capture system audio. After you allow Steno in System Settings, future system-audio recordings work without repeating setup.
No. The recording is saved locally and transcribed on your Mac. No meeting content is uploaded unless you have configured an optional cloud summarization model. The app does send anonymous usage analytics by default (no meeting content), which you can turn off in **Settings → Advanced**. See [How on-device processing works](/privacy/how-on-device-works).
# How to record a Microsoft Teams meeting on a Mac
Source: https://docs.stenoai.co/guides/record-teams-on-mac
Record and transcribe Microsoft Teams calls on macOS without a bot. Steno captures system audio and your microphone locally and writes the notes on-device.
To record a Microsoft Teams meeting on a Mac, use Steno: it captures your Mac's system audio (the other participants) and your microphone together, then transcribes the call entirely on your device and summarizes it locally by default. No bot joins the meeting, and no meeting content is uploaded unless you configure a cloud summarization model. Because it records the audio your Mac plays, it works with Teams in the desktop app or a browser.
## Steps
1. Install Steno and grant microphone and system-audio permission, and turn on the **Record system audio** toggle (see [Installation](/getting-started/installation) and [Recording](/features/recording)).
2. Join your Teams call as usual.
3. In Steno, click **New note** to start recording. It captures your microphone and system audio together.
4. When the call ends, click **Stop**. Steno transcribes the recording and generates notes automatically.
## Why there's no bot
Steno does not join your Teams meeting as a participant. It records the audio locally on your Mac, so no one sees a recording bot or gets a notification from Steno. Recording-consent obligations are your responsibility — see [Confidential use cases](/privacy/confidential-use-cases).
## Speaker labels
With system audio enabled, your microphone and the call audio are recorded as separate channels, so the transcript is labelled `[You]` and `[Others]`. See [Speaker labels](/features/speaker-labels).
***
Yes. Steno records your Mac's system audio, so it captures the call whether you use the Teams desktop app or Teams in a browser tab.
No. Recording, transcription, and summarization all run on your Mac. No meeting content is uploaded unless you have configured an optional cloud summarization model. The app does send anonymous usage analytics by default (no meeting content), which you can turn off in **Settings → Advanced**. See [How on-device processing works](/privacy/how-on-device-works).
# How to record a Zoom meeting on a Mac
Source: https://docs.stenoai.co/guides/record-zoom-on-mac
Record and transcribe Zoom calls on macOS without a bot. Steno captures system audio and your microphone locally, then writes the notes on-device.
To record a Zoom meeting on a Mac, use Steno: it captures your Mac's system audio (the other participants) and your microphone at the same time, then transcribes the call entirely on your device and summarizes it locally by default. No bot joins the meeting, and no meeting content is uploaded unless you configure a cloud summarization model. It works the same way for any calling app, because it records the audio your Mac plays — not Zoom specifically.
## Steps
1. Install Steno and grant microphone and system-audio permission, and turn on the **Record system audio** toggle (see [Installation](/getting-started/installation) and [Recording](/features/recording)).
2. Join your Zoom call as usual.
3. In Steno, click **New note** to start recording. It captures your microphone and system audio together.
4. When the call ends, click **Stop**. Steno transcribes the recording and generates notes automatically.
## Why there's no bot
Steno does not join your Zoom call as a participant. It records the audio locally on your Mac, so no one in the meeting sees a recording bot or gets a notification from Steno. Recording-consent obligations are your responsibility — see [Confidential use cases](/privacy/confidential-use-cases).
## Speaker labels
With system audio enabled, your microphone and the call audio are recorded as separate channels, so the transcript is labelled `[You]` and `[Others]`. See [Speaker labels](/features/speaker-labels).
***
Yes. Steno records your Mac's system audio, so it captures the call whether you use the Zoom desktop app or Zoom in a browser tab.
No. Recording, transcription, and summarization all run on your Mac. No meeting content is uploaded unless you have configured an optional cloud summarization model. The app does send anonymous usage analytics by default (no meeting content), which you can turn off in **Settings → Advanced**. See [How on-device processing works](/privacy/how-on-device-works).
# How to transcribe audio locally on a Mac
Source: https://docs.stenoai.co/guides/transcribe-audio-locally-on-mac
Transcribe meetings and recordings entirely on your Mac with Steno — no cloud upload. On-device Parakeet or Whisper models, with automatic language detection.
To transcribe audio locally on a Mac, use Steno: it runs the speech-to-text model on your device, so recordings are transcribed without uploading anything to a server. Steno offers two on-device engines — **Parakeet** (the default, with a live transcript while you record) and **Whisper** (optional, covering 99 languages) — and auto-detects the language.
## How local transcription works
When you stop a recording, Steno transcribes it on your Mac and saves a plain-text transcript under `~/Library/Application Support/stenoai/transcripts/`. With the default Parakeet engine, a live transcript also appears while you record. See [Live transcription](/features/live-transcription) and [Transcription](/features/transcription).
You can also transcribe an audio file you already have, without recording it in Steno: open the recording options popover (the **···** button next to the record button) and choose **Import audio file…**. Steno transcribes and summarizes the imported file the same way as a live recording.
## Choosing an engine
* **Parakeet (default)** — fast on Apple Silicon, with live transcription. Covers 25 European languages, including English.
* **Whisper (optional)** — covers the full set of 99 languages, including Japanese, Chinese, Korean, Hindi, and Arabic. Switch to it in **Settings → Transcribe**.
See [Transcription models](/models/transcription-models) for the full comparison.
## Privacy
Because transcription runs on-device, your audio never leaves your Mac. This makes local transcription a good fit for confidential recordings — see [Confidential use cases](/privacy/confidential-use-cases).
***
Only for the one-time model download during setup. After that, transcription runs fully offline on your Mac.
Yes, if it's an audio file: use **Import audio file…** from the recording options popover. Transcription always runs locally; summarization uses whichever model is configured in **Settings → AI** (local by default, or a cloud model if you've set one up). There is currently no way to re-transcribe a recording already processed by Steno with a different engine -- the engine choice applies going forward, to new recordings and imports.
# Remote Ollama setup
Source: https://docs.stenoai.co/models/remote-ollama
Offload summarization to a more powerful Mac or workstation on your network. Useful if your laptop cannot run larger models comfortably.
By default, Steno runs Ollama on the same machine where you record. If you have a more powerful Mac on your local network -- a desktop, a Mac Studio, or a Mac mini -- you can run Ollama there and point Steno at it.
This keeps summarization local to your network (no cloud involvement) while freeing your laptop's resources.
## Set up Ollama on the remote machine
Download and install Ollama on the remote machine from [ollama.com](https://ollama.com).
```bash theme={null}
ollama pull gemma4:12b-it-qat
```
You can use any model supported by Ollama, including larger ones that would not run on your laptop.
By default, Ollama only listens on `127.0.0.1`. To accept connections from other machines on your network, set the `OLLAMA_HOST` environment variable before starting Ollama:
```bash theme={null}
OLLAMA_HOST=0.0.0.0:11434 ollama serve
```
For a persistent setup on macOS, add `OLLAMA_HOST=0.0.0.0:11434` to a launchd plist or use a tool like [LaunchControl](https://soma-zone.com/LaunchControl/).
On the remote Mac, go to **System Settings → Network** and note the local IP address (e.g., `192.168.1.10`).
## Point Steno at the remote Ollama
1. Open **Settings → AI**
2. Select **Private Server**
3. Enter the remote machine's address: `http://192.168.1.10:11434`
4. Select the model you want to use
Steno will send the transcript to the remote Ollama for summarization. Audio is never transmitted -- only the text transcript is sent over your local network.
## Notes
* Both machines must be on the same network. This does not work over the internet without additional configuration (VPN or tunnel).
* If the remote machine is not reachable, summarization fails with an error rather than falling back to the local model -- switch back to **Local** in Settings if you need to keep working offline.
* The remote Ollama address is saved in your local Steno config and persists across app updates and reinstalls, as long as your `~/Library/Application Support/stenoai/` data is not deleted.
***
Ollama has no authentication by default. Anyone on your local network who can reach the IP and port can send requests to it. For a trusted home or small office network this is usually acceptable. On a shared or corporate network, use a firewall rule to restrict access to specific IPs.
Not directly without additional setup. You would need a VPN or an SSH tunnel. This is outside the scope of Steno itself -- set up the tunnel first, then point Steno at `http://localhost:[forwarded-port]`.
# Summarization models
Source: https://docs.stenoai.co/models/summarization-models
Steno generates meeting summaries using local Ollama models by default, with optional cloud model support. Choose a model based on your hardware, privacy requirements, and the quality of notes you need.
After transcription, Steno passes the transcript to a language model to generate structured notes: a summary paragraph, key topics, and action items. By default, this uses a local model running via [Ollama](https://ollama.com) entirely on your Mac. You can optionally configure a cloud model (OpenAI, Anthropic, AWS Bedrock, or a custom endpoint) — see the section below.
## Available models
| Model | Size | Notes |
| ------------------- | ---------------------------------- | --------------------------------------------- |
| `gemma4:e2b-it-qat` | \~4.3GB (\~6.5GB on Apple Silicon) | Default. Fast, 128K context. |
| `gemma4:e4b-it-qat` | \~6.1GB (\~8.8GB on Apple Silicon) | Higher quality, modest footprint. |
| `qwen3.5:9b` | \~6.6GB | Strong at structured output and action items. |
| `gemma4:12b-it-qat` | \~7.2GB (\~7.7GB on Apple Silicon) | 256K context, best for long meetings. |
| `gpt-oss:20b` | \~14GB | Highest quality, reasoning. |
On Apple Silicon, the three Gemma 4 models are automatically pulled as MLX/NVFP4 builds instead of the base GGUF -- a meaningful speed win, at a larger download size (shown above).
`gemma4:e2b-it-qat` (Gemma 4 E2B) is the default and handles most meetings well. `llama3.2:3b` is still available for existing users pinned to it, but is deprecated in favor of the Gemma 4 lineup and hidden from the default model picker. Larger models produce better-structured, more detailed notes; the trade-off is processing time and disk space.
## How to change models
1. Open **Settings → AI**
2. Select a model from the list
3. If the model is not yet downloaded, Steno will download it via Ollama
The new model is used for all future recordings. You can re-summarize an existing recording from its detail view using any model.
## Choosing the right model
**For daily use:** `gemma4:e2b-it-qat` (default) works well for most meetings. Notes are clear and concise without requiring significant processing time.
**For important meetings:** `qwen3.5:9b` or `gemma4:12b-it-qat` produce more detailed notes with better-identified action items and more accurate key topics. `gemma4:12b-it-qat` offers a 256K context window, which helps with long meetings.
**For the highest quality:** `gpt-oss:20b` produces the best local output, at the cost of more disk space and slower processing.
## Using a cloud model
Steno optionally supports OpenAI, Anthropic, AWS Bedrock, or a custom API endpoint as an alternative to a local model. If configured, your transcript and any typed notes (not your audio) are sent to that API for summarization.
To configure a cloud model:
1. Open **Settings → AI**
2. Select **Cloud API**
3. Choose the provider and enter your credentials
Cloud models are off by default. If you work with confidential recordings, use a local model.
***
`qwen3.5:9b` is particularly good at identifying and formatting action items from meeting transcripts. If structured output is important, it is worth the extra size over the default `gemma4:e2b-it-qat`.
The default setup (Parakeet + `gemma4:e2b-it-qat`) requires approximately 7GB. The largest model (`gpt-oss:20b`) requires \~14GB. Ollama models are stored in `~/.ollama/models/`.
Yes. Run `ollama rm [model-name]` in Terminal to remove a model and free the disk space. You can re-download it later from within Steno.
Stick with `gemma4:e2b-it-qat` (the default) -- it has the smallest footprint of the curated lineup. The larger models need more RAM to hold the model weights, so on an 8GB Mac they are more likely to cause swapping, which slows both summarization and the rest of your system.
# Transcription models
Source: https://docs.stenoai.co/models/transcription-models
Steno transcribes on-device with one of two engines: Parakeet (the default) and Whisper (optional). Choose the engine that matches your language and workflow needs.
Steno transcribes recordings entirely on your Mac -- no audio is sent to any server. There are two transcription engines: **Parakeet**, the default, and **Whisper**, an optional alternate you can switch to in Settings.
## The two engines
| Engine | Model | Size | Notes |
| ---------------------- | ------------------------------------ | ------- | ------------------------------------------------------------------------- |
| **Parakeet** (default) | `mlx-community/parakeet-tdt-0.6b-v3` | \~572MB | Live transcript while recording; 25 European languages, including English |
| **Whisper** (optional) | `large-v3-turbo` | \~1.6GB | Full 99-language coverage; transcribes after you stop |
**Parakeet** is the default on new installs. It runs via MLX on Apple Silicon and shows a live transcript as you record.
**Whisper** is a single model, `large-v3-turbo`. Choose it in **Settings → Transcribe** when you need a language Parakeet does not cover. Whisper recordings are transcribed only after you stop the recording.
Both engines run 100% on-device.
## How to change engines
1. Open **Settings** in Steno
2. Select **Transcribe**
3. Choose **Parakeet** or **Whisper**
4. If the engine's model is not yet downloaded, Steno will prompt you to download it
The engine is applied to all future recordings. Existing recordings are not re-transcribed automatically -- transcription only runs once, when the recording is first processed.
## Speed
Transcription runs on-device and is fast on Apple Silicon. Parakeet additionally shows a live transcript while you record, so notes are ready shortly after you stop. Whisper transcribes after the recording ends rather than during it.
## Language support
Language coverage depends on the engine you choose.
* **Parakeet (default)** transcribes 25 European languages, including English, French, German, Spanish, Italian, Portuguese, Dutch, Russian, and Polish. A curated subset of six (English, French, German, Spanish, Dutch, Portuguese) can be pinned as the summary output language.
* **Whisper (optional)** covers the full set of **99 languages**, auto-detected. Switch to Whisper for Japanese, Chinese, Korean, Hindi, Arabic, and any language outside Parakeet's set.
Steno auto-detects the language from the audio -- you do not need to configure this manually.
## Accuracy tips
* **Record in a quiet environment** -- background noise is the biggest accuracy factor
* **Use system audio capture** -- capturing both call sides is more accurate than microphone-only for virtual meetings
* **Switch to Whisper for non-European languages** -- Parakeet covers 25 European languages, including English; Whisper covers the full 99-language set
***
Start with **Parakeet** (the default). It runs on-device, shows a live transcript while you record, and covers 25 European languages, including English. Switch to **Whisper** in **Settings → Transcribe** if you need a language outside Parakeet's set.
Yes. Change the engine in **Settings → Transcribe** -- it applies to recordings from that point on. There is currently no way to re-transcribe an already-processed recording with a different engine; you can regenerate its summary, but not its transcript.
Yes. Parakeet covers 25 European languages, including English, French, German, Spanish, Italian, Dutch, Portuguese, Russian, and Polish. For other languages -- including Japanese, Chinese, Korean, Hindi, and Arabic -- switch to the Whisper engine, which covers all 99 languages.
# Pricing
Source: https://docs.stenoai.co/pricing
Steno is free and open source (MIT), with no usage restrictions and no per-seat fees or accounts. A paid managed Enterprise deployment tier is a separate, optional offering.
# Pricing — Steno
Steno is **free and open source** (MIT licensed) -- the MIT license places no
restriction on who can use it or at what scale, including enterprise and commercial
use. There are no per-seat fees, subscriptions, or accounts. Separately, Steno also
offers a paid **Enterprise** tier: managed deployment with admin controls and
organization-specific features, for teams that want that on top of the free app.
## Free — for everyone, including enterprise use
* Price: \$0
* License: MIT (open source) -- no usage restrictions, including for commercial or enterprise deployments
* Source code: [https://github.com/ruzin/stenoai](https://github.com/ruzin/stenoai)
* Account required: no
* Includes: recording (system audio + microphone), on-device transcription (Parakeet and
Whisper engines), on-device summarization (local models via Ollama), live transcription,
report templates, ask-your-meetings chat, speaker labels, and Apple Shortcuts automation.
## Enterprise (optional paid tier)
* A separate, optional offering on top of the free app: managed deployment with admin controls
and organization-specific features.
* Enterprise pricing and packaging is being finalized — details coming soon.
## Optional costs (not paid to Steno)
* **Cloud summarization (optional, off by default):** if you choose OpenAI, Anthropic, or AWS
Bedrock instead of the local model, you pay that provider directly for API usage. Steno does
not resell or mark up these services. A custom endpoint may be a paid API (same as above) or
your own self-hosted server, which may have no per-request cost.
* Everything works with no cloud provider — the default pipeline is fully local and free.
## Platform
* macOS (macOS 14.4 Sonoma or later; Apple Silicon, M1 or later -- Intel Macs are not supported since v0.4.0) and Windows 10/11 (x64) in alpha.
## Download
* [https://stenoai.co/#download](https://stenoai.co/#download)
# Confidential use cases
Source: https://docs.stenoai.co/privacy/confidential-use-cases
Steno is used by healthcare, legal, and finance professionals who cannot send meeting audio to cloud services. Here is how it fits each context.
Steno was built for professionals who work with audio they cannot send to a cloud service. In its default configuration, all transcription and summarization runs on-device, making it suitable for contexts where confidentiality is a hard requirement. If you use the optional cloud summarization feature (OpenAI, Anthropic, AWS Bedrock, a custom endpoint, or an organization adapter), transcripts and any typed notes are sent to the configured provider — do not enable this for confidential recordings.
Steno is not certified under HIPAA, SOC 2, or any other compliance framework. It does not make compliance claims. If you work in a regulated environment, consult your compliance team before using any recording tool -- including Steno. The value Steno provides is technical: in its default configuration, no meeting content (audio, transcript, notes, or summary) leaves your device. The app does send anonymous product-usage analytics (event names, error types, no meeting content) by default -- turn this off in **Settings → Advanced** if your policy requires zero network activity. Do not configure a cloud summarization model for recordings containing regulated or privileged information.
## Healthcare
**The problem with cloud recorders in clinical settings**
Clinical conversations contain protected health information. Cloud-based meeting recorders -- including dedicated AI note-takers -- transmit audio or transcripts to external servers for processing. This creates data transfer obligations and risk, regardless of the vendor's compliance certifications.
**How Steno fits**
* Audio is processed entirely on the clinician's Mac using a local transcription engine
* Transcripts and summaries are stored in the local filesystem, not synced to any cloud
* No vendor relationship is created for PHI processing
* Clinicians can use Steno on their organization-issued Mac without installing third-party cloud services
**Typical uses**
* Recording and summarizing clinical consultations for notes
* MDT (multidisciplinary team) meeting notes
* Supervision sessions, case reviews, and CPD recordings
## Legal
**The challenge for legal professionals**
Legal professional privilege protects communications between lawyers and clients. Routing those conversations through a cloud AI service raises questions about whether the privilege is maintained and whether client consent is required.
**How Steno fits**
* No audio leaves the device -- the privilege analysis is the same as a local dictation recorder
* Transcripts are stored as plain Markdown files you control entirely
* Can be run on a device managed by your firm without creating any external data processor relationship
**Typical uses**
* Client call notes
* Deposition prep and review
* Strategy meeting notes
## Finance
**Why finance teams use local AI tools**
Investment discussions, earnings calls, and client advisory conversations may be subject to information barriers, insider trading rules, or client confidentiality obligations. Cloud recording services create an external data trail.
**How Steno fits**
* Processing is isolated to the analyst's or advisor's local machine
* No audio or transcript is transmitted externally (unless a cloud model is configured -- off by default)
* Compatible with firm policies that prohibit cloud services for sensitive discussions
**Typical uses**
* Investment committee notes
* Client relationship meeting notes
* Research call notes
## General guidance
For any regulated or confidential context:
1. Use the local model (the default) -- do not configure a cloud model for sensitive recordings
2. Verify audio is not being captured by any other app or system service before recording
3. Store recordings on an encrypted volume (macOS FileVault covers this)
4. Review your organization's recording consent policies -- Steno does not manage consent
5. Steno saves raw audio in `~/Library/Application Support/stenoai/recordings/` -- delete files you no longer need
***
No. Steno does not notify other participants that a recording is being made, manage consent workflows, or add disclosure notices. Consent obligations vary by jurisdiction and context -- this is your responsibility.
Steno is a standard macOS app distributed as a signed DMG. It can be deployed via MDM (e.g., Jamf) by distributing the `.app` bundle. Contact the project via GitHub for enterprise deployment questions.
# How on-device processing works
Source: https://docs.stenoai.co/privacy/how-on-device-works
Steno transcribes and summarizes your meetings using AI models that run entirely on your Mac. No audio is sent to any server.
Steno processes your recordings using two AI models that run locally on your hardware. Your audio, transcripts, and summaries stay on your Mac unless you optionally configure a cloud summarization model (see below) -- meeting content is never uploaded by default. The app does send anonymous product-usage analytics (see below), which is on by default and can be turned off in Settings.
## The processing pipeline
When you stop a recording, Steno runs this sequence entirely on your device:
```
Audio file (local)
→ Parakeet or Whisper (on-device transcription)
→ transcript.txt (saved locally)
→ Ollama (on-device LLM)
→ summary.md (saved locally)
```
In the default local configuration, no step in this pipeline contacts an external server. The models are downloaded once during setup and then run offline indefinitely. If a cloud summarization model is configured, the transcript step is sent to that provider's API instead of local Ollama.
## What runs on your Mac
**Transcription -- Parakeet or Whisper**
Steno transcribes with one of two on-device engines. **Parakeet** (the default) runs via MLX on Apple Silicon and produces a live transcript while you record. **Whisper** (`large-v3-turbo`) is an optional alternate for the full 99-language set, selectable in **Settings → Transcribe**. Both run entirely in-process -- the model weights are stored under `~/Library/Application Support/stenoai/` and there is no transcription API call.
**Summarization -- Ollama**
Steno bundles [Ollama](https://ollama.com), which manages and runs small language models locally. When Steno starts, it launches a local Ollama server that by default listens only on `127.0.0.1` (loopback). All summarization and query requests go to `http://localhost:11434` -- they do not leave your machine unless `OLLAMA_HOST` is overridden in your environment.
## What network requests does Steno make?
| Request | When | Why |
| ------------------------- | --------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------- |
| Model download | First launch / setup only | Downloads transcription and Ollama model weights |
| Update check | On app launch, then periodically | Checks `github.com/ruzin/stenoai` for new releases |
| Anonymous usage analytics | On events like recording start/stop, transcription/summarization completion, and errors | Anonymous product-usage telemetry (event names and counts only -- never meeting content); on by default, can be turned off in **Settings → Advanced** |
None of these requests include your audio, transcript, or summary content. To verify this yourself, you can monitor network traffic with [Little Snitch](https://www.obdev.at/products/littlesnitch/) or macOS's built-in `nettop` while Steno is processing a recording.
## Where your data is stored
All data lives in `~/Library/Application Support/stenoai/`:
```
stenoai/
├── recordings/ # .webm audio files
├── transcripts/ # .txt verbatim transcripts
├── output/ # .md summaries and notes
└── models/ # transcription model weights
```
Ollama model weights are stored separately in `~/.ollama/models/` (Ollama's default location).
## Using a cloud model (optional)
Steno optionally supports OpenAI, Anthropic, AWS Bedrock, or a custom API endpoint ("Cloud API" in Settings) as an alternative to the local Ollama model, plus a "Private Server" mode that points at a remote Ollama instance you control. If you configure one of these in **Settings → AI**, your transcript and any notes you've typed for that meeting (not audio) are sent to that provider's API for summarization.
Organizations can also connect Steno to a self-hosted adapter under **Settings → Organisation**, which proxies summarization requests server-side. As with the other cloud options, this sends the transcript (and notes) off-device but never the audio.
Cloud and organization modes are opt-in, off by default, and clearly indicated in the UI. If you work with confidential data, use the local model.
***
Yes, anonymous product-usage analytics are on by default -- event names like "recording started" or "summarization completed," plus error types and an anonymous install ID. This never includes audio, transcripts, notes, or summaries. You can turn it off in **Settings → Advanced**. Separately, the only other outbound requests from the app are the update check and model downloads during setup. If you enable a cloud summarization model, transcripts are sent to that provider's API.
Yes, after the initial setup. Download the app and models on a connected machine, then move the installation to your air-gapped environment. All processing runs offline.
Steno is designed for professionals handling confidential audio. Because all processing is local and no meeting content (audio, transcript, notes, or summary) leaves your device by default, it is suitable for use cases where sending audio to a cloud service would be inappropriate -- clinical consultations, legal calls, financial briefings. If you need zero network activity at all, turn off anonymous usage analytics in **Settings → Advanced**. Review your organization's specific policies before using any recording tool for regulated data.
Steno bundles a copy of Ollama and starts it as a local process on `127.0.0.1:11434` when the app launches. It is not accessible from the network. When you quit Steno, Ollama is stopped.