Skip to content
@jamjet-labs

JamJet

Agents that don't lose control. Safety layer for AI agents: policy, audit, approval, recovery, cost, memory. Works with any framework via MCP/A2A.
README.md

JamJet Labs

JamJet is the action-control plane for AI agents. One policy file. One audit trail. Across hooks, guardrails, MCP gateways, SDKs, and custom runtimes.

Write the safety policy once. Run it everywhere your agents can act.

Flagship repos

  • jamjet — Python SDK + runtime + policy engine. Start here.
  • jamjet-policy — Portable policy adapters: Claude Code hook, MCP shim, OpenAI guardrail, CLI. One policy YAML, six runtimes.
  • engram — Durable memory for AI agents. Independent product.

JVM

Protocols

  • jamjet-a2a — Agent-to-Agent protocol SDK in Rust.

What's coming next

The portable policy layer just shipped. Next: deeper Cloud sync for approvals + audit (web UI for jamjet approve), Java/Spring adapter, MCP HTTP/SSE transport.

Links

Pinned Loading

  1. jamjet-examples jamjet-examples Public

    Older JamJet examples. Primary examples now live in jamjet-labs/jamjet under examples/.

    Python 1

  2. jamjet-benchmarks jamjet-benchmarks Public

    Older JamJet benchmark experiments. Phase 1 evidence now lives in jamjet-labs/jamjet under evidence/.

    Python

  3. jamjet jamjet Public

    The open-source safety layer for AI agents — block unsafe tool calls, require approval, enforce budgets, audit, replay.

    Rust 11 5

Repositories

Showing 10 of 13 repositories
View all repositories

Top languages

Loading…

Most used topics

Loading…