| What it does | Push-to-talk voice input for coding with AI
| Run up to 16 AI coding agents side by side, let your tests pick the winner, and undo any step
|
|---|
| Stage | Startup
| Not given
|
|---|
| Founded | 2026
| Not given
|
|---|
| Based in | Not given
| Not given
|
|---|
| Pricing model | Open source
| Freemium
|
|---|
| Pricing | Free, open source (AGPL-3.0-only)
| Three plans: Starter $17/month or $170/year (saves $34 yearly), Pro $29/month or $290/year (saves $58 yearly), Ultra $50/month or $500/year (saves $100 yearly). First monthly Starter or Pro plan includes 5-day trial for $1 (₹99 in India). Pro includes 1,500 minutes cloud voice monthly and Video Stud
|
|---|
| Key features | - Push-to-talk dictation in most desktop applications
- On-device speech recognition with no audio uploaded
- Works offline after model download
- Support for 29 languages including 25 European languages, Mandarin Chinese, Hindi, Korean and Japanese
- Local meeting recording, transcription and evidence-linked notes
- Rebindable push-to-talk hotkeys for Windows and macOS
- Text insertion into active field in most desktop applications
- Free and open-source desktop app under AGPL-3.0-only
| - Agent Grid: up to 16 agents in real terminals side by side
- Undo for agents: rewind any step byte for byte
- Point and Fix: click a bug in the browser and say the fix
- Token Guardian: budgets that stop runaway agent runs
- Overnight Autopilot: queue work to run unattended with morning review
- SwarmPane Voice: dictation and voice commands on Mac and Windows
- Video Studio: generate launch videos from descriptions
- Signed approvals: nothing edits files without your approval
|
|---|
| What makes it different | - Speech recognition runs entirely on your own machine with no audio or transcript uploaded
- Works completely offline after initial model download
- Completely free and open source with no payment, account or activation required
- Types directly into active application without copying from dictation box
- Supports 29 languages with multiple on-device recognition models
| - Run up to 16 coding agents in parallel, each in its own copy of your project
- Undo any agent step exactly, byte for byte
- Your own tests automatically judge which agent's work is best
- Works with CLIs you already use: Claude Code, Codex, Cursor, Gemini and others
- Voice dictation to edit code hands-free
|
|---|
| Free plan | Not given
| Not given
|
|---|
| Open source | Not given
| Not given
|
|---|
| Platforms | Not given
| Not given
|
|---|
| Public API | Not given
| Not given
|
|---|
| Community votes | 0 | 0 |
|---|
| DR (Domain Rating by Ahrefs) | 8 ● no change since last check | 4 ● no change since last check |
|---|
| Trust Flow (Majestic) | 4 | 7 |
|---|
| Citation Flow (Majestic) | 21 | 26 |
|---|
| Referring domains (Majestic) | 48 | 39 |
|---|
VocalCode
SwarmPane