Run opencode on Mingles Router
opencode builds providers on the Vercel AI SDK, so an OpenAI-compatible endpoint is a few lines of JSON and no plugin.
Pick your model and paste your key — every command and value on this page updates live.
Endpoint: https://router.mingles.ai/v1 · key stays in your browser ·
no key entered? commands show <your-key>.
Add the provider
1. Put the key in the environment
opencode resolves {env:NAME} at load, so the config file stays shareable:
export MINGLES_API_KEY="<your-key>"2. Add the block to opencode.json
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"mingles": {
"npm": "@ai-sdk/openai-compatible",
"name": "Mingles Router",
"options": {
"baseURL": "https://router.mingles.ai/v1",
"apiKey": "{env:MINGLES_API_KEY}"
},
"models": {
"deepseek-ai/DeepSeek-V4-Flash-0731": {
"name": "deepseek-ai/DeepSeek-V4-Flash-0731",
"limit": { "context": 128000, "output": 8192 }
}
}
}
}
}3. Select it
Run /models in opencode and the provider is in the list.
Exact values
| options.baseURL | https://router.mingles.ai/v1 |
| options.apiKey | <your-key> |
| model id | deepseek-ai/DeepSeek-V4-Flash-0731 |
| limit.context | 128000 |
Switch model
The models map takes as many entries as you like — one line each, switchable with /models mid-session.
Before you file a bug: read the limits
Reasoning models spend the output budget on internal thinking, output is capped at 8192 tokens, there is no KV cache, and there are no built-in web tools. Most “it broke on Mingles Router” reports are one of these. See Model limits & behavior →
Frequently asked
Why @ai-sdk/openai-compatible and not @ai-sdk/openai? +
The openai-compatible package makes no assumptions about OpenAI-only features, which is what you want against a third-party endpoint. It is the package opencode’s own docs use for this case.