​
1. Overview

OpenCode is an open-source AI coding assistant supporting 75+ models and local deployment. With Nebula API, you can use mainstream and latest models (e.g. GPT, Claude, Gemini) in OpenCode and configure custom providers and models. Download: https://opencode.ai/

​
2. Quick setup (Nebula API)

​
2.1 Get an API key

Create and copy your API key from the Nebula API console.

​
2.2 Configure a custom provider in OpenCode

  1. Open OpenCode and go to Server / Provider settings.
  2. Add a custom provider (Configure an OpenAI compatible provider).
  3. Fill in:
    • Provider ID: e.g. nebula (lowercase, digits, hyphens or underscores)
    • Display name: e.g. Nebula API
    • Base URL: https://llm.ai-nebula.com/v1 (must end with /v1)
    • API key: paste your Nebula API key
  4. Under Models, add the models you need (e.g. gpt-4o, claude-sonnet-4-20250514). Model IDs must match the Nebula model list.
  5. After saving, use providerID/modelID in the model selector (e.g. nebula/gpt-4o).
OpenCode custom provider setup Configuration reference (matching the screenshots above): OpenCode provider and API config OpenCode model config

​
2.3 Switching models

In the chat or settings, select the configured provider and model (e.g. nebula/gpt-4o) to switch. OpenCode switch model
  • For self-hosted or backup Nebula, set Base URL to that address, e.g. http://your-server:3003/v1.
  • Default model can be set in project or global opencode.json with "model": "nebula/modelID".

​
3. Some models require the Responses API (important)

Some models use a different API than standard Chat Completions and require the Responses API. If you see an error like this when using such a model in OpenCode: Some Azure / OpenAI models only support the Responses API and do not support the traditional Chat Completions endpoint. Example error:
The chatCompletion operation does not work with the specified model, gpt-5.1-codex. Please choose different model and try again.
The screenshot below shows how this error appears in the UI: Error when using gpt-5.1-codex with Chat Completions This means the request was sent as Chat Completions while the model only exposes Responses API on the server. You need to configure OpenCode to use the correct API.

​
3.1 Models that require the Responses API (typical)

Model ID / familyNotes
gpt-5.1-codexGPT 5.1 Codex, coding; Responses API only
gpt-5.2-codexGPT 5.2 Codex, same as above
o3 / o3-proReasoning models; some deployments only offer Responses API
o4-mini and other o4Same; check deployment and vendor docs
computer-use-previewExperimental model used with Responses API computer-use tools
For the full and up-to-date list, see Azure / vendor documentation, e.g.: You can make OpenCode use the Responses API for a given model by adding the apiCompatibility parameter for that model in the config file. No code changes are required. 1. Locate the config file
  • Windows: C:\Users\<username>\.config\opencode\opencode.jsonc
  • macOS / Linux: ~/.config/opencode/opencode.jsonc
    You can also use opencode.json or opencode.jsonc in your project directory.
2. Add the parameter to the model config under your custom provider In your existing provider (e.g. custom provider pointing to Nebula or your own API), add "apiCompatibility": "responses" for models that need the Responses API (e.g. gpt-5.1-codex). Example: Responses API requirement for some models opencode.jsonc: apiCompatibility responses for gpt-5.1-codex 
{
  "$schema": "https://opencode.ai/config.json",
  "disabled_providers": [
    "backup_nebula"
  ],
  "provider": {
    "backup_nebula": {
      "name": "backup_nebula",
      "npm": "@ai-sdk/openai-compatible",
      "models": {
        "gpt-5.2": {
          "name": "gpt-5.2"
        },
        "gpt-5.1-codex": {
          "name": "gpt-5.1-codex",
          "apiCompatibility": "responses"
        },
        "gpt-5.1-chat": {
          "name": "gpt-5.1-chat"
        }
      },
      "options": {
        "baseURL": "xxxxxxxx"
      }
    },
    "backup_v1": {
      "name": "backup_v1",
      "npm": "@ai-sdk/openai-compatible",
      "models": {
        "gpt-5.2": {
          "name": "gpt-5.2"
        },
        "gpt-5.1-codex": {
          "name": "gpt-5.1-codex",
          "apiCompatibility": "responses"
        },
        "gpt-5.1-chat": {
          "name": "gpt-5.1-chat"
        }
      },
      "options": {
        "baseURL": "http://34.87.38.188:3003/v1"
      }
    }
  }
}
After saving, when you select that provider’s gpt-5.1-codex in OpenCode (e.g. backup_nebula/gpt-5.1-codex), requests will be sent to that baseURL in Responses API format. 3. Success example The screenshot below shows a normal chat in OpenCode using a correctly configured model (e.g. backup_v1/gpt-5.1-chat). OpenCode chat after successful config

​
3.3 Summary

ScenarioWhat to do
Models that only support Responses (e.g. gpt-5.1-codex, gpt-5.2-codex)In opencode.jsonc, under the custom provider’s models, add "apiCompatibility": "responses" for that model (see config example and screenshots above).
Regular models (e.g. gpt-4o, claude-sonnet)No apiCompatibility needed; use Nebula or a custom provider as in “Quick setup” above.
CategoryModel ID examplesNotes
Coding / Codexgpt-5.1-codexEnsure Responses API (see Section 3)
General chat / codinggpt-5.2, gpt-5.1-chatUse as-is; no Responses config
Long context / generalclaude-sonnet-4-20250514, claude-sonnet-4-5Same
Cost-effectivegpt-4.1-mini, gemini-2.5-flashSame
See the Nebula model list for available models and IDs.

​
Success example

The screenshot below shows a normal chat in OpenCode using Nebula API. OpenCode success with Nebula API

​
5. References