Two interviews.
A prototype to review.
Follow the Harbor research interviews from project review problems to a prototype, with a next step assigned to Maya.
Free to use · Private by design
Turn your recordings into
private knowledge. For free.
Transcribe, summarize, and build knowledge from your recordings.
Your recordings and notes stay on your computer. No subscription.
Free to use No subscription Early preview
0104
What you can do
Start with a recording.
Here’s what you get.
Read the transcript, check the notes, and export the result. Use Memory and MCP when you need to work across recordings.
Transcription
Import audio or video. Get a transcript with timestamps.
Speaker labels
Separate speakers, review their labels, and assign names.
Summaries & tasks
Generate key points, action items, chapters, and mind maps.
Meeting knowledge
Review facts from multiple recordings and track what changed.
File exports
Save notes, transcripts, and subtitles as files you can reuse.
MCP access
Let an AI client search saved text in approved workspaces.
01 / Private AI transcription
A transcript you can check against the audio.
Import a meeting, interview, or voice memo. Transcription runs on your computer, with the recording available for playback.
- Timestamps
- Jump to a segment and replay the original words.
- Speaker labels
- See separate speaker turns and correct the names.
- Local workspaces
- Keep related recordings and their results together.
02 / Summarize & organize
Summary, action items, chapters, and a mind map.
Generate meeting notes from the transcript. Review the output alongside the source before sharing it.
- Summary & key points
- Read a short account of the conversation with source timestamps.
- Action items
- Review suggested next steps, owners, and due dates.
- Chapters & mind maps
- Browse the conversation by topic or as a visual outline.
03 / Meeting knowledge · Preview
Track decisions across recordings.
Select recordings in a workspace to extract facts. Review new information, keep earlier decisions in History, and ask questions with transcript evidence.
- Multiple sources
- Build a shared set of facts from selected conversations.
- Changes & history
- Review an updated owner or decision alongside the earlier version.
- Cited answers
- Open the transcript excerpt behind a fact or answer.
Preview limit: 20 selected recordings and 60 facts per workspace.
Build a meeting knowledge base across recordings
Keep a copy
Export the files you need.
Use your notes in a document, add subtitles to a video, or keep structured data for another tool.
- Markdown
- Formatted notes
- TXT
- Plain text
- SRT
- Timed subtitles
- JSON
- Structured results
04 / MCP
Let your AI search approved recordings.
Connect a compatible AI client to read saved notes and transcripts through the local MCP server.
- Choose the workspaces
- Limit each connection to the recordings you approve.
- Search and read saved text
- Retrieve recording details, notes, and transcript segments. Access is read-only.
- Review and revoke access
- Check the local access history and revoke a connection when needed.
MCP shares saved text, not raw audio or structured Memory. External clients may send that text to their model provider.
See MCP tools and access controls
Local processing
Download the models.
Then work offline.
Once the required models are installed, transcription, speaker processing, and summaries run on your computer. Recordings and generated results stay in your local workspace.
Connecting an external AI client through MCP is optional and has a separate data boundary.
See where your data goes at each stepYOUR DEVICELOCAL
Your recording
TranscriptSpeakersNotes
No automatic cloud fallback.
Try the examples
Three scenarios.
Six sample recordings.
Listen to the interviews and follow-ups. Compare the transcripts, summaries, and next steps.
Two interviews.
A release plan.
A change of owner.
Compare the kickoff and update. Check who owns the release and which date the team agreed to.
A discovery call.
An agreed pilot scope.
Review the customer’s requirements and the follow-up that narrows the pilot to what the team will deliver.
The price is simple
Free. The whole workflow.
Free local transcription, speaker labels, summaries, tasks, and mind maps.
No subscription. No per-minute fees. No paid feature tier.
$0ZoneVoice desktopJoin the free preview ↗
A few details
Good questions.
Does my audio leave my computer?
The desktop app runs transcription, speaker separation, voiceprint matching, and summaries locally. There is no automatic cloud fallback for processing your recordings.
Is ZoneVoice free?
Yes. ZoneVoice is free software to use, with no subscription, per-minute charges, or paid feature tier. It is currently in early preview, so access is by invitation while we refine the desktop experience.
How does speaker recognition work?
ZoneVoice separates speakers and collects candidate voices in a local library. You confirm who a voice belongs to. Later recordings can suggest or match known speakers, and your manual corrections take priority.
Is Memory included in the desktop preview?
Yes, as a desktop preview. Select up to 20 completed recordings per workspace to extract up to 60 facts, review changes, preserve history, and ask questions with transcript evidence. MCP reads saved notes and transcripts; Memory facts and history are available inside ZoneVoice.
What can I take out of ZoneVoice?
Export your notes, transcripts, subtitles, and structured data as Markdown, text, SRT, or JSON. Your source recordings remain in your local workspace.
When can I try it?
The free desktop preview is available for macOS and Windows by invitation. Join the waitlist with your email, and we’ll notify you when an invitation is available.
ZoneVoice desktop