| What it does | Codebase intelligence platform for AI agents and teams
| Point at anything on your Mac and speak to give your coding agent a brief with screenshots
|
|---|
| Stage | Startup
| Not given
|
|---|
| Founded | 2024
| Not given
|
|---|
| Based in | Not given
| IN
|
|---|
| Pricing model | Freemium
| Not given
|
|---|
| Pricing | Free: 2 public repos, 10 AI answers/month. Pro: $15/month (moving to $20/month in November, or $16/month yearly), includes private repos, auto-sync on every push, security scanning, $5/month AI credits. Teams: $20/seat/month with 3-seat minimum, includes 25 shared repos, shared $5/seat/month credit
| Free: two hours of monthly transcription included. Pro (monthly): $6.99/month for ten hours of transcription. Pro (yearly): $69/year for ten hours of transcription. Pro (lifetime): $169 one-time for ten hours of transcription.
|
|---|
| Key features | - Code health scoring with defect prediction
- Blast radius and change impact detection
- Auto-generated documentation rebuilt on every commit
- Git intelligence: hotspots, owners and coupling detection
- Architecture and decision tracking
- MCP tools for Claude Code and Cursor
- PR bot with change review and test detection
- Security scanning: CVEs, secrets and reachability
- Dead code detection and unused exports
- Change Records with policy enforcement
| - Point and speak to create briefs with annotated screenshots
- Works with Claude Code, Cursor, Codex, Windsurf, ChatGPT, Claude.ai and more
- Mixed-language support in single sentence
- Three customizable persona templates for brief format
- Offline operation with fallback to on-device processing
- Automatic credential detection and withholding
- One-line command installation
- No account or sign-up required
|
|---|
| What makes it different | - Open source under AGPL-3.0, self-hostable with no LLM calls required to build the index
- Every published benchmark includes sample size, corpus and statistical methods
- Code health scores trained on years of real commits and defect fixes, with 0.74 AUC prediction accuracy
- Eliminates need to switch tools: integrates with Claude Code, Cursor, VS Code, GitHub and team dashboards
| - Everything stays on your Mac—no data is kept on servers, and screenshots never leave unless you send them to an agent yourself
- Automatic task memory means your agent remembers context from previous conversations about the same issue
- Fully open source under MIT licence so you can verify all privacy and security claims
- Supports mixed-language voice input in a single sentence
|
|---|
| 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) | 11 ▲ +1 (up) since last check | Not checked yet ● no change since last check |
| Trust Flow (Majestic) | 8 | 3 |
| Citation Flow (Majestic) | 34 | 25 |
| Referring domains (Majestic) | 244 | 38 |