Skip to content

Zed

Zed's agent panel can use any OpenAI-compatible provider. Add Inferray in your settings and its models show up in the panel's model picker.

Add the provider

Open your settings (Zed → Settings → Open Settings, or zed: open settings from the command palette) and add:

{
  "language_models": {
    "openai_compatible": {
      "Inferray": {
        "api_url": "https://api.inferray.com/v1",
        "available_models": [
          {
            "name": "claude-sonnet-4-6",
            "display_name": "Claude Sonnet 4.6",
            "max_tokens": 1000000,
            "max_output_tokens": 64000,
            "capabilities": { "tools": true, "images": true, "parallel_tool_calls": false, "prompt_cache_key": false }
          },
          {
            "name": "gpt-5.5",
            "display_name": "GPT-5.5",
            "max_tokens": 1050000,
            "max_output_tokens": 128000,
            "capabilities": { "tools": true, "images": true, "parallel_tool_calls": true, "prompt_cache_key": false }
          }
        ]
      }
    }
  }
}

name is the model ID from the catalogue; max_tokens is its context length.

Add the key

Open the agent panel's settings (agent: open settings), find Inferray under the providers, and paste your Inferray key. Zed stores it in your system keychain, not in the settings file.

Use it

Open the agent panel, pick an Inferray model, and start a thread. Requests appear in your logs.

Models that work well here