Developer Tools
opencodex
Source: lidge-jun/opencodex · ★ 4675 · TypeScript
Last updated 2026-07-23 · benchmark measured 2026-10-06 — deterministic & reproducible
Local proxy that routes OpenAI Codex and Claude Code through any LLM provider.
opencodex at a glance
opencodex is a self-hosted local proxy that lets developers run OpenAI Codex and Claude Code against more than 40 different LLM providers without changing their client setup.
Good for
- Developers who use Codex or Claude Code but want to run requests through a different LLM provider
- Multi-provider LLM users who want one proxy to route to Anthropic, Google, xAI, Ollama, or OpenAI-compatible endpoints
- AI engineers who need streaming, tool calls, reasoning, and image inputs to work across providers without patching the client
Questions people ask
What does opencodex actually do?
It is a local proxy that translates Codex and Claude Code requests so they can be routed to over 40 LLM providers, including Anthropic, Google, xAI, Ollama, and OpenAI-compatible endpoints.
Do I need to modify Codex or Claude Code to use it?
No, opencodex works without requiring any changes to the client.
How do I install and start opencodex?
You install it with 'npm install -g @bitkyc08/opencodex' and then run 'ocx start' to start the proxy; a desktop app (beta) is also available for macOS, Windows, and Linux.
What authentication methods does opencodex support?
It supports API keys, OAuth logins (with auto-refresh) for providers like xAI, Anthropic, and Kimi, and ChatGPT account pooling where new sessions auto-route to the lowest-usage healthy account.
Does opencodex work with Claude Code as well as Codex?
Yes, the command 'ocx claude' launches Claude Code against the same port, with every routed model available in the /model picker and support for auto-context up to 1M and sub-agents.
Written by Legit.Show from opencodex.me on 2026-10-10. It does not change the score.
Are you the maker of opencodex? Claim this listing. It is free, takes a meta tag, a DNS record or GitHub admin rights, and never changes the score.
Launched something? Add your product, free.
To cite this score: legit.show/s/gh-lidge-jun-opencodex/2026-10-06. That address never changes; this page moves with every re-measure.
Legit.Show scored opencodex 72/100 on 2026-10-06, measured across all 7 frames from its public surface. legit.show/s/gh-lidge-jun-opencodex/2026-10-06
Is opencodex production-ready?
Legit.Show scores opencodex 72 out of 100 — the simple average of its 7 measured frames. Legit.Show ran its deterministic 7-Frame production-readiness benchmark on opencodex (public-surface assessment), measured from the public surface with no LLM in the scoring path. Its strongest frame is Reliability; its weakest is Privacy. Every frame it averages is published with its evidence on the Legit.Show listing.
The 7 Frames
- Performance — 78/100
- Accessibility — 91/100
- Security — 25/100How this compares across the catalogue: The Web Security Baseline
- Privacy — 25/100How this compares across the catalogue: The Privacy Gap
- Reliability — 100/100
- Standards — 92/100
- Discoverability — 90/100
What we measured
- No Content-Security-Policy and no HSTS.
- Served over HTTPS with a valid certificate.
- Real Lighthouse performance run — 162 ms to first byte.
- Returns a proper 404 for unknown routes.
- 3 of 3 sampled routes reachable.
- No privacy policy found.
- Sets cookies / loads scripts with no consent prompt.
- Discoverable: structured data, OpenGraph image, canonical URL.
Who it's for
Developers using Codex · Claude Code users · Multi-provider LLM users · AI engineers
Sources and updates
- Description
- Taken from opencodex.me's own website on 2026-07-23.
- Code
- github.com/lidge-jun/opencodex.
- Benchmark
- Measured by Legit.Show from the live site, the way any visitor sees it — with no access to its code or accounts. Same method for every product, and no AI decides the score. Last checked 2026-10-06.
Put together from public information, without opencodex's involvement. If anything here is wrong, tell us and a person will check it.
Visit opencodex → · Alternatives to opencodex → · How this was measured →