What is ACP?
ACP is a governance and auditing platform designed for teams using AI coding agents. It monitors and logs every tool call, agent delegation, and step in agent chains, giving you visibility and control over what your AI assistants are doing. The tool integrates with popular AI coding environments including Claude Code, Cursor, CrewAI, and LangGraph. Setup is minimal; a single command gets you started with logging and governance across your agent workflows. This is particularly useful for teams building production AI systems where you need to track decisions, debug agent behaviour, and maintain compliance records.
Key Features
Tool call logging
Records every function call and action taken by your agents
Agent chain tracking
Visualises and monitors the sequence of delegations between agents
Audit trails
Creates a complete record of decisions and outputs for compliance and debugging
Multi-framework support
Works with Claude Code, Cursor, CrewAI, LangGraph, and similar platforms
Quick setup
Single command initialisation without extensive configuration
Freemium access
Free tier available for individuals and smaller projects
Pros & Cons
Advantages
- Straightforward implementation with minimal configuration required
- Broad framework compatibility covers most popular agent platforms
- Clear audit trails help with debugging complex agent behaviour
- Freemium model lets you try it without upfront cost
Limitations
- Limited information available about advanced governance features or customisation options
- Unclear how pricing scales for larger teams or enterprise deployments
- May add latency to agent operations depending on logging overhead
Use Cases
Debugging multi-agent workflows to understand where decisions went wrong
Maintaining compliance records for regulated industries using AI systems
Monitoring agent behaviour in production to catch unexpected actions
Team review of agent outputs before deployment or user-facing releases
Understanding agent decision paths for safety and transparency audits