Client work
Record the detail, not just the decision.
Turn client calls, interviews and briefings into searchable notes without exporting the raw recording.
Transcribe.
Privately.
100% on-device. Zero cloud. Zero subscriptions. whisper.cpp transcription with speaker diarization and local AI summaries.
Free 7-day trial · No account required · macOS 14+ on Apple Silicon
Keep the meeting. Keep the context. Keep it on your Mac.
LocalWhisper is for work where a transcript is useful, but sending the recording to another cloud is not an option.
Research
Find the line you need later.
Search speakers, timestamps and action items instead of replaying long recordings.
Internal teams
Share outcomes, not sensitive audio.
Create a summary and action list locally, then decide exactly what leaves the machine.
Seamless
Integration
LocalWhisper notices when a meeting starts in Zoom, Teams, or Meet — with real microphone activity, not just an open window — and offers to record and transcribe it. Locally.
Raw Telemetry
Data Uplink
0 BYTES UPLOADED
Processing Locality
100% ON-DEVICE
Throughput
1 HR ≈ 1 MIN
Engine
WHISPER.CPP · METAL
Recording Active
Absolute Clarity
Speaker attribution mapped with on-device diarization. Timestamped. Readable, high-contrast typography with names — never color alone.
14:02:11
Sarah Chen [Host]
"Alright, let's review the deployment pipeline for Q3. We need to ensure zero downtime during the database migration. Have we finalized the rollback strategy?"
14:02:45
Alex Vance [Engineer]
"Yes. The staging environment mirrored production exactly. We ran the test suite three times this morning. Latency remained under acceptable thresholds."
14:03:10
Sarah Chen [Host]
"Excellent. Push the commit to the main branch at 0200 hours. I'll monitor the telemetry."
Insight,
Not Uplink
Summaries and action items generated by the LLM you choose — Ollama, any OpenAI-compatible endpoint, or the built-in offline summarizer. The transcript never leaves the machine.
Generated locally · Stayed on your Mac
Q3 Deployment Sync — Summary
Team confirmed the staging environment mirrors production. Rollback strategy approved. Database migration scheduled with zero-downtime window.
Action items
- 01Push the commit to main at 02:00.
- 02Monitor telemetry during rollout.
- 03Verify rollback checkpoints after migration.
Voice Studio
Clone your voice. Offline.
Record a 30-second sample and generate speech in your own voice with Qwen3-TTS on Apple Silicon. The model, the sample, and every generated clip stay on disk — yours.
System-wide dictation
Speak into any app.
One global hotkey. Talk, release, and the transcribed text lands wherever your cursor is — editor, browser, terminal. Processed by the same local engine.
Transcript Chat · New
Ask your meetings anything.
Open a recording and interrogate it — decisions, action items, who promised what. Attach several meetings to one conversation. Answers come from the LLM you choose; hour-long transcripts are condensed on-device before a single token leaves the app.
Podcast Studio · New
Two hosts. Your meeting. Zero cloud.
Turn any transcript into a 5–30 minute two-host discussion — a curious host and an expert, scripted by your LLM and voiced by local TTS, including your own cloned voice. Stitched into an episode you can play, read, and export.
Own It
Free
£0
Forever
- Cloud transcription with your own API key
- 7-day full trial of the local pipeline
- No account required
One-time
Pro
£29.99
One-time · Yours forever
- Unlimited local transcription — whisper.cpp on Metal
- Speaker diarization + local AI summaries
- Voice Studio cloning + system-wide dictation
- Transcript chat + podcast generation
- Free updates via Sparkle
Already have a license? Activate in Settings → License. · Store