codex-claude-bridge

Provides a bidirectional bridge between Claude Code (via Claude Code Channels from an MCP server) and OpenAI Codex CLI (via an MCP server exposing blocking tool calls). It routes Claude- and Codex-initiated messages through the bridge and renders a real-time web UI while the two agents discuss.

Evaluated Mar 30, 2026 (0d ago)
Repo ↗ Automation mcp agent-to-agent claude-code codex-cli channels web-ui typescript bun
⚙ Agent Friendliness
48
/ 100
Can an agent use this?
🔒 Security
24
/ 100
Is it safe for agents?
⚡ Reliability
22
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
55
Documentation
70
Error Messages
0
Auth Simplicity
85
Rate Limits
0

🔒 Security

TLS Enforcement
20
Auth Strength
20
Scope Granularity
0
Dep. Hygiene
35
Secret Handling
50

Bridge is intended for localhost. No auth/authorization for the local web UI/internal endpoints is documented, and no TLS requirements are described (assumed plain localhost). The README does not discuss secret handling within the bridge. Dependency hygiene cannot be fully assessed from provided metadata (only @modelcontextprotocol/sdk is listed).

⚡ Reliability

Uptime/SLA
0
Version Stability
25
Breaking Changes
30
Error Recovery
35
AF Security Reliability

Best When

You run both Claude Code and Codex on the same machine and want a developer-local bridge with interactive visibility (web UI) and mostly real-time Codex->Claude turns.

Avoid When

You need strong security guarantees, public network exposure, or strict symmetric full-duplex push between both agents.

Use Cases

  • Let Claude Code and Codex CLI collaborate on coding discussions in a shared thread
  • Enable live, near-real-time A2A-style interaction by using Claude Channels plus Codex MCP tool polling
  • Human-observed multi-agent coding conversation via the local web UI

Not For

  • Production-grade deployment across machines/tenants (bridge is documented as localhost-only)
  • Security-sensitive environments without additional hardening (no auth details for the local web UI/internal endpoints are provided)
  • Use cases requiring fully symmetric push notifications in both directions

Interface

REST API
Yes
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: Uses local stdio MCP configuration (no auth documented for the bridge itself)
OAuth: No Scopes: No

README specifies Codex side needs an OpenAI API key or ChatGPT login, but does not document authentication/authorization for the bridge’s local HTTP/web UI or internal endpoints.

Pricing

Free tier: No
Requires CC: No

Open-source bridge; costs depend on underlying Claude Code and Codex/OpenAI usage.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Claude-initiated messages are queued and require Codex to poll or issue another request; Codex->Claude is smoother
  • Claude must include reply_to when replying to Codex; otherwise routing back to Codex may fail (though UI may still show the reply)
  • Both agents need to be on the same machine (localhost bridge)
  • Requires Claude Code research preview Channels and the --dangerously-load-development-channels flag

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for codex-claude-bridge.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

6533
Packages Evaluated
19870
Need Evaluation
586
Need Re-evaluation
Community Powered