| What it does | Security monitoring for AI coding tools that detects secrets and PII before they're sent
| MCP server audits, builds and research
|
|---|
| Stage | Startup
| Startup
|
|---|
| Founded | Not given
| 2026
|
|---|
| Based in | Not given
| US
|
|---|
| Pricing model | Not given
| Paid
|
|---|
| Pricing | Free individual app, forever. Team plans start at 10 seats with annual contracts; pricing quoted in writing per deployment. Enterprise plans available for large fleets.
| Write-Safety Audit: $2,500 fixed for one server with up to 40 tools. Write-path rebuild: $6,000 fixed in total for one existing or generated MCP server with up to 40 tools. New MCP server: from $7,500 fixed in the blueprint. MCP blueprint: $1,000, part of the build price if you sign the build within
|
|---|
| Key features | - Real-time warning as you type
- Supports Cursor, Claude Desktop, VS Code with Copilot, ChatGPT Desktop, and Windsurf
- Deterministic detection with no AI model
- Local detection with no network call
- Manager dashboard with category, severity, tool, and engineer information
- End-to-end encryption with manager-held keypair
- Per-organization hash chain audit trail
- macOS 13+ on Apple Silicon
| - Write-path testing for lost updates, stale writes, silent no-ops, type coercion, partial multi-step writes, and annotation honesty
- Control-checked findings verified against vendor APIs
- Security review of permission boundaries, read-only switches, and credentials in model context
- Tool catalog review including token weight and annotation flags
- Evidence packs independently timestamped and verifiable locally
- Fix pull requests for high-severity findings
- Before-and-after agent benchmarks on two task sets
- 60-minute audit walkthrough and 60-day defect warranty
- MCP server builds to ten rules covering sign-in, tools, reads, writes and tests
- AI-assisted audits with separate refuter model and person-reproduced findings
|
|---|
| What makes it different | - Secrets are masked on the device before any record is created, so live secrets never reach company servers
- Deterministic regex rules instead of AI models, making detection explainable and consistent offline
- Contractual guarantee that detection data cannot be used for performance reviews
| - One-person company: the person who scopes your work is the person who runs it
- Every finding verified three ways: control-checked, refuter-challenged, and person-reproduced
- Public findings filed upstream; track record of issues and pull requests fully visible
- Fixed prices with no padding: clean reports are a result, finding more earns nothing extra
- Transparency on AI use and testing methodology in every deliverable
|
|---|
| 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 ▲ +7 (up) since last check | 3 ● no change since last check |
| Trust Flow (Majestic) | 3 | 0 |
| Citation Flow (Majestic) | 24 | 22 |
| Referring domains (Majestic) | 70 | 18 |