opencode-gpt-image-plugin

An opencode plugin that adds image generation with OpenAI's GPT image models through any OpenAI-compatible / CLIProxyAPI endpoint.

It registers one tool your agents can call:

| Tool | Purpose | Models | | --- | --- | --- | | gpt_image | Still images | gpt-image-2 (default), gpt-image-1.5 |

The agent picks the specific model via the model argument, guided by per-model hints baked into the tool schema, and can set size, quality, background, and output_format.

For adult/explicit image generation, use the companion opencode-grok-imagine-plugin; GPT image models refuse explicit content upstream.

Why a plugin (and not model config)

opencode's model config only drives the chat/completions interface. GPT image generation uses a different endpoint (/v1/images/generations) and returns base64 image data to write to disk. A plugin tool is the supported way to reach it and save results.

Install

npm install -D opencode-gpt-image-plugin

Then reference it in opencode.json:

{
  "$schema": "https://opencode.ai/config.json",
  "plugin": ["opencode-gpt-image-plugin"]
}

Restart opencode (plugins load at startup).

Configure

Configuration is read from plugin options first, then environment variables.

Option A — environment variables

export GPT_IMAGE_BASE_URL="https://your-proxy.example.com"   # or CPA_BASE_URL
export GPT_IMAGE_API_KEY="sk-..."                            # or CPA_API_KEY
# Only for self-signed LAN proxies (see TLS below):
# export GPT_IMAGE_INSECURE_TLS=1

opencode.json:

{
  "plugin": ["opencode-gpt-image-plugin"]
}

Option B — plugin options (tuple form)

{
  "plugin": [
    ["opencode-gpt-image-plugin", {
      "baseURL": "https://your-proxy.example.com",
      "apiKey": "{env:MY_KEY}",
      "insecureTLS": false
    }]
  ]
}

Config matrix

| Setting | Plugin option | Env var(s) | Default | | --- | --- | --- | --- | | Base URL | baseURL | GPT_IMAGE_BASE_URL, CPA_BASE_URL | — (required) | | API key | apiKey | GPT_IMAGE_API_KEY, CPA_API_KEY | — (required) | | Skip TLS verify | insecureTLS | GPT_IMAGE_INSECURE_TLS | false |

The base URL may include or omit a trailing /v1; the plugin normalizes it.

TLS

TLS certificates are verified by default. Only disable verification for a self-signed LAN proxy you control, via insecureTLS: true or GPT_IMAGE_INSECURE_TLS=1.

Usage

Just ask an agent:

• "generate an image of a minimalist mountain logo on white" → gpt_image
• "make a 1024x1536 transparent-background icon of a rocket" → gpt_image

Where files are saved

• Project assets — pass output_dir (e.g. assets/img, public/images).
• Scratch / experiments — omit output_dir. Files go to a self-ignored <project>/.gpt-tmp/ folder.

gpt_image arguments

| Arg | Type | Notes | | --- | --- | --- | | prompt | string | required | | model | enum | gpt-image-2 | gpt-image-1.5 | | count | 1–4 | number of images | | size | enum | auto, 1024x1024, 1536x1024, 1024x1536 | | quality | enum | auto, low, medium, high | | background | enum | auto, transparent, opaque | | output_format | enum | png, jpeg, webp | | output_dir | string | asset path, or omit for scratch |

Depending on the upstream proxy, some parameters (notably exact size) may be approximated rather than honored precisely.

Response & endpoint

• POST /v1/images/generations → { data: [{ b64_json }], size, quality, ... }

Notes

• Model availability depends on your upstream account/proxy. The plugin surfaces the upstream error text when a model or parameter isn't available.
• Content policy is enforced upstream by OpenAI, independent of this plugin.

License

MIT © Lu Cao