Developer Guide
Everything you need to integrate OpenBox into your agent workflows.
Integrations
Most integrations below wrap a runtime agent framework, a deployed workflow, graph, or crew. Claude Code is different: it's governed via its hooks system as a shift-left dev-tooling integration, not a runtime framework adapter. See Getting Started for the full distinction.
Claude Code
Hook configuration, event model, and observe/enforce modes for governing Claude Code dev sessions.
CrewAI
Governance for multi-agent crews and collaborative workflows. Every agent action is tracked automatically.
CopilotKit
Run OpenBox above CopilotKit; today the documented agent-framework bridge is LangGraph.
Cursor coming soon
Governance for Cursor IDE agents via hooks. Every prompt, shell command, MCP call, and file read is governed.
Deep Agents
Per-subagent governance for DeepAgents workflows. Every nested call is captured automatically.
LangChain
Python middleware governance for LangChain agents, model calls, tools, and hook-level telemetry.
LangGraph
Governance for graph-based, stateful agent workflows. Every node and state transition is recorded.
Mastra
Governance for TypeScript AI agents and tool calls. Your existing Mastra code stays unchanged.
n8n
Govern your AI Agent node with one node swap. Your Chat Model, Memory, and Tool connections stay unchanged.
OpenRouter
Governance for OpenRouter agents, with attested evidence of which upstream provider served each prompt.
OpenClaw coming soon
Tool governance and LLM guardrails for OpenClaw agents. Every tool call is evaluated against your policies.
Temporal
Add the OpenBox plugin to your Temporal worker. Your existing workflows and activities stay unchanged.
Shared Reference
| Guide | Description |
|---|---|
| Working with llms.txt | Machine-readable documentation for LLMs and AI tools |