Pi | EasyRouter API Docs

Pi (pi-mono) tutorial — A minimal, highly extensible terminal coding framework. Add EasyRouter as a custom provider via its config file.

Project Introduction

Pi (pi-mono) is a minimal, highly extensible terminal coding framework. It ships with TypeScript extensions, skills, prompt templates, themes, a tree-structured session model, and 15+ built-in providers. Adding EasyRouter as a custom OpenAI-compatible provider is a single config edit away.

GitHub: https://github.com/can1357/pi-mono

📦 Prerequisites

What you'll need

Pi (pi-mono) installed
An EasyRouter endpoint (use https://easyrouter.io/v1 as usual)
An EasyRouter API key (generated in the console)
A model ID exactly matching one exposed by your EasyRouter console (e.g. claude-sonnet-4-6, gemini-2.5-flash)

🔑 Step 1: Set the API key env var

export EASYROUTER_API_KEY="your-easyrouter-api-key"

🔧 Step 2: Edit the Pi config file

Add EasyRouter as a provider in Pi's config file:

{
  "providers": {
    "easyrouter": {
      "baseUrl": "https://easyrouter.io/v1",
      "api": "openai-completions",
      "apiKey": "$EASYROUTER_API_KEY",
      "models": [
        {
          "id": "claude-sonnet-4-6",
          "name": "Claude Sonnet 4.6",
          "contextWindow": 200000,
          "maxTokens": 128000,
          "input": ["text"],
          "reasoning": true,
          "compat": {
            "requiresReasoningContentOnAssistantMessages": true,
            "thinkingFormat": "anthropic",
            "reasoningEffortMap": {
              "minimal": "low",
              "low": "low",
              "medium": "medium",
              "high": "high",
              "xhigh": "max"
            }
          }
        },
        {
          "id": "gemini-2.5-flash",
          "name": "Gemini 2.5 Flash",
          "contextWindow": 1000000,
          "maxTokens": 65536,
          "input": ["text"],
          "reasoning": false
        }
      ]
    }
  }
}

Field notes

baseUrl uses the standard EasyRouter endpoint https://easyrouter.io/v1
api is always openai-completions
apiKey uses the $EASYROUTER_API_KEY variable reference
compat.requiresReasoningContentOnAssistantMessages: true keeps reasoning content on historical messages when thinking mode is active
reasoningEffortMap maps Pi's effort levels to the target model's supported levels; you can omit this block for non-reasoning models

✅ Step 3: Launch & verify

Launch Pi, pick easyrouter in the provider selector, then pick a model. Send a test message — a normal response means you're set.

❓ Troubleshooting

Issue	Fix
easyrouter not in the provider list	Check the config file path and JSON format
401 Unauthorized	Make sure `EASYROUTER_API_KEY` is exported in the current shell
Reasoning model acts weird	Verify `requiresReasoningContentOnAssistantMessages` and `thinkingFormat` match the model
Model not found	Confirm `models[].id` matches your EasyRouter console

Pi

📦 Prerequisites

🔑 Step 1: Set the API key env var

🔧 Step 2: Edit the Pi config file

✅ Step 3: Launch & verify

❓ Troubleshooting

On this page