Use Codex with Claude, Gemini or a Chat Completions-only relay#

ThinkWatch Lite connects Codex to a gateway on the same computer that accepts the OpenAI Responses API, the only format Codex uses with a custom provider. The gateway converts each request and its answer between Responses and the upstream’s format: Anthropic Messages for Claude, Google Gemini, or OpenAI Chat Completions for a relay that offers nothing else. The model is named in Codex’s configuration or set by a routing rule.

Before you start#

• ThinkWatch Lite, installed.
• Codex, either the Codex CLI or the Codex in the ChatGPT desktop app, run at least once.
• An API key for Anthropic, Google Gemini or the relay.

Steps#

1. On the Upstreams page, choose New upstream and pick the service on the first step, Service, which goes straight to Connection: Anthropic for Claude, or Google Gemini; each fills in the address and protocol. For a relay, choose its service under Platforms and relays, or Custom for any other. OpenRouter fills in its address and protocol. For Sub2API, New API / One API and Custom, enter the relay’s Base URL without an endpoint path such as /chat/completions, and set Protocol to OpenAI Chat Completions; for GLM, for example, choose Custom with https://api.z.ai/api/paas/v4, or …/api/coding/paas/v4 on a GLM Coding Plan. A base URL that ends with its own version, such as /v4 or Volcengine Ark’s /api/v3, is used as written from ThinkWatch Lite 2026.10.5. Enter the API key and choose Next, which checks the connection. If the relay lists no models, or leaves some out, type each missing model ID in the box at the end of Models and press Enter. Choose Next, then Create.

2. On the Clients page, choose Connect… on the Codex row. The dialog shows the change to ~/.codex/config.toml:

   Field	Value
   model_provider	thinkwatch
   model_providers.thinkwatch.base_url	The gateway address with /v1, by default http://127.0.0.1:8788/v1
   model_providers.thinkwatch.wire_api	responses
   model_providers.thinkwatch.experimental_bearer_token	A new key named codex
   model_providers.thinkwatch.http_headers	X-ThinkWatch-Client = "codex"
   name, requires_openai_auth, supports_websockets in the same table	ThinkWatch, false, false

   Choose Connect, then reopen the terminal. The ChatGPT desktop app reads the same file and picks up the change after a restart.

3. Name the model. Connecting leaves model as it was, so Codex keeps asking for the model it used before. Either:

   • set model = "<model ID>" at the top of ~/.codex/config.toml, before the first [section], or pass -c model=<model ID> for a single run; or
   • keep Codex’s model and rewrite it: on the Routing page, open the default route and choose Add rule, add the conditions Model gpt-* and Key codex, set On match to Forward and Forward to to the upstream, and under Parameter rewrites enter the model ID in Change model to. Choose Add, then Save.

Notes#

• Codex’s model table. Codex carries metadata for its own models, such as the context window, inside the program. A model it does not know, such as a Claude or Gemini model, runs on fallback metadata with a 272,000-token context window, and Codex warns: “Model metadata for <model> not found. Defaulting to fallback metadata; this can degrade performance and cause issues.” For a model with a smaller window, model_context_window in config.toml sets the window Codex assumes. The gateway’s model list does not appear in Codex’s model picker, as Codex expects a catalog in its own format.
• Conversion. Requests and streamed answers are converted in both directions. Traffic marks such requests Converted; fields the target format cannot carry are dropped and listed in the request details. Server-side tools such as web search run only at the provider they belong to and are dropped in conversion. Codex accepts only responses for wire_api, so this conversion is what makes a Chat Completions-only relay usable.
• Tools, history and compaction. From ThinkWatch Lite 2026.10.11, every tool Codex declares reaches the upstream, including those it lists in its input rather than in tools, and tool calls come back under the names Codex gave them. Local shell calls, tool search and reasoning-effort changes in the history are converted as well. Compacting a long session works: the upstream writes a summary, which Codex keeps as its compaction and sends back in later requests. A compaction encrypted by OpenAI cannot be read by another upstream, and a request that carries one is refused with an error saying so. Instructions Codex adds in the middle of a conversation stay where they are, so the system prompt stays the same from turn to turn and the prompt cache keeps hitting.
• Credentials. With requires_openai_auth = false, Codex authenticates to the gateway with its own key and sends no OpenAI key or ChatGPT token. A ChatGPT account signed in from the app can still serve the OpenAI models next to Claude or Gemini: each request goes to an upstream that lists the model asked for.
• Sessions. Codex lists sessions started before and after connecting separately. codex resume <session ID> -c model_provider=thinkwatch continues an earlier session through the gateway. After Restore…, sessions started while connected can still be opened, and go straight to OpenAI.
• Cost. A rewritten request is priced as the model actually sent. Codex marks no prompt-cache breakpoints, so on a route to an upstream in Anthropic Messages format, or to a Claude model on Bedrock that AWS lists for prompt caching (Claude 3.5 Sonnet v2, Claude 3.7 Sonnet, and Claude 4.5 and later), the gateway marks the tools, the system prompt and the last two user turns, and each turn reads what the previous one cached. Older Claude models on Bedrock, such as Sonnet 4, get no marks, and an upstream that refuses them receives the request again without them. Cache writes and reads are priced at the price sheet’s cache rates; at Anthropic, writes cost 25% more than input and reads a tenth of it.

Related: Features, Use Claude Code with GLM, DeepSeek or Kimi, Install and update.