CC Code Conduit

// the missing layer between your team and AI

Security & governance for AI coding agents like Claude Code and Codex

Code Conduit is a self-hosted proxy that sits between your developers and their AI coding tools. Monitor every prompt, audit every tool call, block dangerous commands before they execute, and enforce content policies — all with one environment variable. Works with Claude Code, OpenAI Codex, OpenClaw, and any Anthropic- or OpenAI-compatible client.

Sign up free → Sign in ›

What Code Conduit does

  • Claude Code audit logging

    Every prompt, response, tool call, and cost is captured and searchable. Full session replay for incident review and compliance.

  • Block Claude Code tool calls

    Real-time policy engine inspects every tool call mid-stream. Dangerous Bash commands, file writes, and network calls are blocked before they execute.

  • Codex prompt monitoring

    OpenAI Codex requests get the same parsing, redaction, and policy enforcement as Anthropic traffic. One proxy for both providers.

  • AI coding agent governance

    Per-team policies, secret redaction, prompt profiles, and per-user access keys. Built for security and platform teams running AI at scale.

  • Self-hosted Claude Code proxy

    Deploy to your own infrastructure or use the hosted version. No SDK, no code changes — just point ANTHROPIC_BASE_URL at your tenant.

  • Cost & usage tracking

    Per-user, per-team, and per-model spend dashboards. Set budgets, alert on overruns, and export usage for chargeback.

Setup in 5 minutes

One environment variable enables Code Conduit for Claude Code, Codex, or any compatible AI coding agent. No SDK to install, no code changes, no measurable added latency. 

export ANTHROPIC_BASE_URL=https://your-org.codeconduit.ai
claude "refactor the auth module"
# ✓ logged · ✓ policy check passed · ✓ streaming · $0.034