Home / Integrations / OpenAI Codex

Use Intercept with OpenAI Codex

Add policy enforcement to Codex in one line.

Config file: ~/.codex/config.toml (global) or .codex/config.toml (project)

Codex uses TOML format, not JSON. Each server is a [mcp_servers.name] table.

A standard MCP server in your Codex config -- no controls.

config.toml
[mcp_servers.stripe]
command = "npx"
args = ["-y", "@stripe/mcp-server"]

[mcp_servers.stripe.env]
STRIPE_SECRET_KEY = "sk_live_..."

Wrap the server command with Intercept. Same config structure, policy enforced.

config.toml
[mcp_servers.stripe]
command = "npx"
args = ["-y", "@policylayer/intercept", "-c", "policy.yaml", "--", "npx", "-y", "@stripe/mcp-server"]

[mcp_servers.stripe.env]
STRIPE_SECRET_KEY = "sk_live_..."
Rate limits
Cap tool calls per minute, hour, or day
Access controls
Allow, deny, or conditionally gate any tool
Spend caps
Limit financial operations with daily thresholds
Audit logs
Full trace of every tool call and policy decision
policy.yaml
version: "1"
default: deny

tools:
  list_customers:
    rules:
      - action: allow
        rate_limit: 30/minute

  create_refund:
    rules:
      - name: "daily cap"
        rate_limit: 10/day

  delete_repository:
    rules:
      - action: deny
npx -y @policylayer/intercept
github.com/policylayer/intercept →
// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.