Use Nanobot with Manifest
Nanobot can route through custom OpenAI-compatible providers from
~/.nanobot/config.json. Use Manifest in the
built-in custom provider slot and point
Nanobot's default model at auto.
9 models routed through Manifest over the last 30 days.
Why use Nanobot with Manifest
Personal agents can burn through tokens while searching, planning, and editing. Manifest lets Nanobot keep one endpoint while routing requests to lower-cost models whenever the task does not need a premium frontier model.
- Use one auto model in Nanobot
- Mix API keys, subscriptions, local models, and custom providers
- Fallback automatically when providers rate limit or fail
- Track usage and cost per Nanobot agent in the Manifest dashboard
- Set spending caps before long autonomous runs
- Self-host Manifest for local-first routing
Setup in 60 seconds
Update agents.defaults and add the
custom provider block below to Nanobot's
config file.
Merge this JSON into ~/.nanobot/config.json.
{
"agents": {
"defaults": {
"provider": "custom",
"model": "auto"
}
},
"providers": {
"custom": {
"apiKey": "mnfst_YOUR_KEY",
"apiBase": "https://api.manifest.build/v1"
}
}
}custom
https://api.manifest.build/v1
mnfst_YOUR_KEY
auto
Nanobot tokens processed on Manifest
Models in use
# Model |
Auth type | Tokens / last 30 days |
|---|---|---|
1.Owl Alpha | API Key | 106 M |
2.Glm 5.2 | Subscription | 42 M |
3.Nex N2 Pro | API Key | 22 M |
4.Gpt 5.5 | Subscription | 1 M |
5.Nemotron 3 Ultra 550b A55b | API Key | 527 K |
6.Minimax M3 | Subscription | 512 K |
7.Qwen3.7 Max | Subscription | 216 K |
8.Gemini 3.1 Pro Preview | Subscription | 37 K |
9.Claude Sonnet 4 6 | Subscription | 2 K |