AirGapScribe
Summary
Cloud transcription services are a non-starter for legal depositions, therapy sessions, or financial calls — the moment audio leaves the device, compliance conversations begin and liability follows. AirGapScribe runs Whisper locally on Windows, transcribes without touching a network, and in its paid tier generates meeting minutes, follow-up drafts, and action items on-device.
The core workflow is a Windows tray app: record from microphone, system audio, or both; transcription runs on-device using Whisper models; export as .txt. The free demo caps sessions at ten minutes and watermarks exports — enough to verify the local pipeline, not enough for production use. The paid tier removes those caps, adds an on-device AI assistant that queries across saved transcripts, generates structured deliverables, and accepts .txt, .md, .pdf, or .docx files as context. NVIDIA GPU acceleration is supported and cuts generation time significantly, but CPU-only machines still work. There is no API, so AirGapScribe does not plug into existing pipelines — you pull deliverables out manually.
Bottom line: The right call for a legal or clinical professional who needs to transcribe hour-long sessions on a locked-down Windows machine without touching a cloud service; the wrong call for any team that needs transcription output flowing automatically into a CRM, ticketing system, or downstream automation.
Pricing Plans
Subscription- Price
- 5.99 USD/month or 49 USD/year
- Free Tier
- 10 min per session, smaller models, watermarked exports, limited vocabulary
Demo
Free forever with limits
- 10 min sessions
- tiny/base/small models
- watermarked .txt export
Pro
Unlimited sessions and full features
- Unlimited sessions
- Local AI assistant
- All models including large-v3
- Clean exports and projects
View full pricing on airgapscribe.com →
Pricing may have changed since last verified. Check the official site for current plans.
Community Performance Report Card
No community ratings yet. Be the first to rate this tool!
Pros
Sign in to edit- All audio and transcript data stays on the local machine with zero bytes uploaded, so legal and clinical professionals meet data-residency requirements without a separate compliance review of a third-party vendor.
- The on-device AI assistant queries across multiple saved meeting transcripts simultaneously, which means you can surface a decision buried in a client's third session without manually reviewing each file.
- Context documents (.txt, .md, .pdf, .docx) attach directly to the assistant session, so generated minutes and follow-up drafts reference your own briefing materials rather than hallucinating from transcript text alone.
- NVIDIA GPU acceleration is detected and offered automatically, cutting local AI generation from minutes to seconds — which makes the tool practical for back-to-back meeting workflows rather than an overnight batch job.
- Structured .json export in the paid tier means transcripts can be parsed and archived programmatically, even without an API, so teams with light scripting capacity can build a manual ingestion step into an existing file workflow.
Cons
Sign in to edit- The free demo caps recordings at ten minutes and watermarks all exports, which means you cannot test transcription accuracy on a real deposition or client call without paying first — teams evaluating fit for compliance-sensitive use cases are essentially buying blind.
- Custom vocabulary (proper names, acronyms, domain jargon) is locked behind the paid tier, so a medical practice or law firm cannot assess whether Whisper handles their terminology before committing.
- There is no API — transcripts and generated deliverables leave the app only by manual export, so any team whose workflow depends on transcription feeding a CRM, ticketing system, or data pipeline will hit this wall immediately and move to a self-hosted Whisper deployment or a service like AssemblyAI that exposes an endpoint.
- Windows-only distribution means teams with macOS or Linux users have no path to adoption — a mixed-OS legal or clinical team cannot standardize on this tool without maintaining a separate transcription workflow for non-Windows machines.
About
- Platforms
- Windows
- API Available
- No
- Self-Hosted
- Yes
- Last Updated
- 2026-09-16T10:24:29.457Z
Best For
Who it's for
- Privacy-focused professionals in legal, health, or finance
- Users needing Windows tray-based one-click recording
- Teams avoiding cloud transcription services
What it does well
- Transcribing confidential legal depositions offline
- Generating meeting minutes and action items without data leaving the device
- Processing audio in air-gapped or restricted networks
Add notes, reviews, and benchmarks so the next visitor gets a clearer picture.
Compare AirGapScribe
Spotted incorrect or missing data? Join our community of contributors.
Sign Up to ContributeFrequently Asked Questions
- Is AirGapScribe free?
- AirGapScribe has a permanent free tier alongside paid upgrades (paid plans from 5.99 USD/month or 49 USD/year). You can keep using a baseline version indefinitely without paying.
- Is AirGapScribe open source?
- No — AirGapScribe is a closed-source tool. Source code is not publicly available.
- Can I self-host AirGapScribe?
- Yes. AirGapScribe supports self-hosting on your own infrastructure.
- What platforms does AirGapScribe support?
- AirGapScribe is available on: Windows.
Best AirGapScribe alternatives →
Curated lists that include this category
Cloud transcription creates compliance headaches
Legal depositions, therapy sessions, and financial calls cannot use cloud services because audio leaving the device triggers compliance reviews and liability questions. AirGapScribe runs Whisper models locally on Windows so no audio or transcripts reach any network.
Core workflow
The Windows tray app records from microphone, system audio, or both. Transcription stays on-device. Exports are .txt files. The free demo limits sessions to ten minutes and adds watermarks. The paid tier removes those limits and adds an on-device AI assistant that queries saved transcripts, generates meeting minutes and action items, and accepts .txt, .md, .pdf, or .docx context files. NVIDIA GPU acceleration speeds generation while CPU-only operation remains possible.
Pricing and limits
Subscription pricing is 5.99 USD per month or 49 USD per year. Free tier limits include ten minutes per session, smaller models, watermarked exports, and restricted vocabulary.
Who it is for / who should skip it
Privacy-focused professionals in legal, health, or finance who need Windows tray recording and must keep all data on-device will find the local pipeline and assistant useful. Teams that require an API or want to test full accuracy and custom vocabulary without paying first should skip it because the demo caps recordings at ten minutes, watermarks exports, and locks vocabulary behind the paid tier.