Production integration docs · examples use router-api.mengbi-ai.net and a personal API key placeholder.
Prepare
Prepare Pi Agent with a personal Mengbi Router key and a current model ID.
Prepare 1
Use the Pi Agent configuration supported by your current installation.
npm install -g --ignore-scripts @earendil-works/pi-coding-agent
pi --versionPrepare 2
- Install the latest stable Pi Agent release.
- Create a personal Mengbi API key in the console.
- Copy the exact model ID from the model catalog.
NoteClient configuration fields can change between versions. Prefer the documented OpenAI-compatible or native provider setting.
Configure
Connect Pi Agent to the endpoint that matches its supported protocol.
Configure 1
Use the Pi Agent configuration supported by your current installation.
{
"providers": {
"mengbi": {
"name": "Mengbi Router",
"baseUrl": "https://router-api.mengbi-ai.net/v1",
"apiKey": "$MENGBI_ROUTER_API_KEY",
"api": "openai-completions",
"models": [
{
"id": "gpt-5.6-sol",
"name": "GPT-5.6 Sol",
"reasoning": true,
"input": ["text"],
"cost": { "input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0 },
"contextWindow": 128000,
"maxTokens": 16384
}
]
}
}
}Configure 2
Use the Pi Agent configuration supported by your current installation.
export MENGBI_ROUTER_API_KEY="sk-mengbi-xxxxxx"
pi --provider mengbi --model gpt-5.6-solConfigure 3
Use the Pi Agent configuration supported by your current installation.
NoteClient configuration fields can change between versions. Prefer the documented OpenAI-compatible or native provider setting.
Verify
Send a small request and check the key, endpoint, and model if it fails.
Verify 1
Use the Pi Agent configuration supported by your current installation.
pi --list-models --provider mengbiVerify 2
Use the Pi Agent configuration supported by your current installation.
- Install the latest stable Pi Agent release.
- Create a personal Mengbi API key in the console.
- Copy the exact model ID from the model catalog.
- Keep the key in a local secret store and never commit it.