PeerAI Trust Center
Product

Product Security

Built secure, shipped secure, kept secure.

Studio's product-security posture spans the build (secure SDLC), the runtime (Tauri sandboxing, sidecar IPC, BYO providers), and the maintenance loop (daily scans, release security gate, responsible disclosure).

a390ee4

Commitments

  • Tauri-native runtime

    Studio runs in a Tauri shell — capability-scoped IPC between WebView and Rust backend, no Electron-style full Node access in the renderer.

  • Sandboxed sidecars

    Python and Rust sidecars run as separate processes; communication via authenticated IPC. Process isolation bounds the impact of any single component compromise.

  • Local-first secrets

    Credentials are stored in OS keychain by default with an encrypted file fallback at ~/.peerai/credentials.json. Path safety is enforced via core/path_safety.py.

    Data security
  • BYO LLM, BYO database

    No PeerAI-side proxy or aggregator. Customer credentials connect directly to customer-selected services.

    AI Security
  • SHA-pinned CI actions

    All GitHub Actions in security-relevant workflows are pinned by SHA, not by tag. Documented in dev-guide and audited per release.

  • Per-release security gate

    Lockfile-focused grype scans on 9 targets must pass before any release tag.

    Continuous monitoring

Evidence

  • Daily vulnerability scans

    grype + pip-audit + cargo-audit + bun pm audit + continuous Aikido.

    Vulnerability rollup
  • CycloneDX SBOM per release

    2,672 components tracked.

    SBOM
  • Responsible disclosure

    48h ack SLA, 7-day assessment, 30-day critical fix target.

    Disclosure