# sybil-solutions/codex-shim

Local Responses-API shim that exposes Factory BYOK models (and optional ChatGPT GPT-5.5 passthrough) to Codex Desktop.

Repository: https://github.com/sybil-solutions/codex-shim
Canonical: https://ross.abutalabs.com/products/codex-shim
Language: Python
License: MIT
License Family: permissive
Last push: 2026-08-25T22:59:52+00:00

## Health v2 (maintenance only)
Score: 58/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 35, longevity 7
- inputs: {"age_days": 103, "days_push": 8, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1064, forks 107 (observed 2026-08-28T04:03:26.619641+00:00)

## What it is
A local Python/aiohttp server that exposes an OpenAI Responses-compatible endpoint on loopback, letting Codex Desktop use any BYOK model (OpenAI, Anthropic, OpenRouter, local proxies) plus optional ChatGPT and Cursor subscription passthrough. It translates streaming responses, tool calls, and reasoning blocks between Codex's expected format and each upstream API without rebuilding Codex.

## Use cases
- use my own API keys with Codex Desktop
- add custom models to the Codex model picker
- route Codex requests to Anthropic or OpenRouter models
- use my ChatGPT subscription's Codex model via API
- route coding tasks to the cheapest capable model automatically
- proxy and rewrite prompts before they hit an LLM upstream

## When to choose
- you use Codex Desktop and want BYOK or subscription models it doesn't officially expose
- you want local model routing without rebuilding or patching Codex
- you need tool-call and streaming translation between Responses API and other LLM APIs
- you want a local proxy layer in front of LLM requests for prompt policy or deduplication

## When to avoid
- you don't use Codex Desktop or Cursor
- you need a production-grade, officially supported API gateway
- you rely on Codex versions whose internal formats the shim hasn't been tested against
- you want a hosted service rather than a local loopback server

## Facets
- artifact type: service
- maturity: active
- function: http-server, middleware, llm-inference, sdk
- domain: large-language-models, developer-tools, apis, backend
- platform: python, windows, cross-platform, cli
- tags: api-shim, responses-api, codex, byok, model-routing, proxy, aiohttp, chatgpt-passthrough, cursor-passthrough, local-server, macos, linux

## Member repositories
- sybil-solutions/codex-shim (main) score 58

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:26.619641+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:56:06.628609+00:00, confidence not recorded.
  - readme: https://github.com/sybil-solutions/codex-shim (fetched 2026-08-28T04:03:26.619641+00:00, sha 26a051cc9449)
- Data as of 2026-08-30T08:39:29.467469+00:00.