Use OpenCode with Manifest
OpenCode can use OpenAI-compatible providers from
~/.config/opencode/opencode.json.
Add Manifest as a provider and keep using OpenCode while Manifest chooses the right model for each coding request.
— models routed through Manifest over the last 30 days.
Why use OpenCode with Manifest
Coding assistants mix simple edits, long-context search, and hard reasoning. Manifest lets OpenCode keep one stable model name while routing those requests across your providers, subscription plans, local models, and fallbacks.
- Use one manifest/auto model in OpenCode
- Route easy edits to cheaper models and hard changes to stronger ones
- Fallback across providers when quota or rate limits hit
- Track token usage, cost, model choice, and failures in one dashboard
- Keep the OpenCode workflow unchanged
- Self-host Manifest or use Manifest Cloud
Setup in 60 seconds
Add Manifest as an OpenAI-compatible provider in your global
opencode.json config and set the default
model to manifest/auto.
Add this block to ~/.config/opencode/opencode.json.
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"manifest": {
"npm": "@ai-sdk/openai-compatible",
"name": "Manifest",
"options": {
"baseURL": "https://api.manifest.build/v1",
"apiKey": "mnfst_YOUR_KEY"
},
"models": {
"auto": {
"name": "Manifest Auto"
}
}
}
},
"model": "manifest/auto"
}https://api.manifest.build/v1
mnfst_YOUR_KEY
manifest/auto
OpenCode tokens processed on Manifest
Models in use
# Model |
Auth type | Tokens / last 30 days |
|---|---|---|
