Overview

Welcome to codingplan.ai documentation. Get started with our unlimited coding plan API.

Key Features

• Unlimited API requests - No per-token pricing, just concurrency lanes
• OpenAI & Anthropic API compatible - Drop-in replacement for existing code
• State-of-the-Art Models - Access models like Kimi K2.5, DeepSeek, and OpenRouter (Pony-Alpha)
• Weighted Voting - Participate in model selection with weights tied to your subscription
• IP whitelisting - Enhanced security with IP-based access control
• Dead Request Detection - Atomic concurrency management with automatic lane release

Get Started

Quick Start

Get up and running in 5 minutes

API Reference

Complete API documentation

Voting System

Learn about community-driven model selection

Tool Integration Guides

Connect your favorite coding tools to codingplan.ai. Our OpenAI-compatible API works with all major AI coding assistants.

IDE Extensions

• Claude Code
• Cline
• Roo Code
• Kilo Code

CLI Tools

• Codex CLI
• Gemini CLI
• Aider
• OpenCode

Frameworks

• OpenAI SDK
• LangChain
• LlamaIndex
• Any OpenAI-compatible client

About CPAI

CPAI is a modern, high-performance API platform that provides unlimited access to state-of-the-art open-source coding models. Instead of paying per token, users subscribe to concurrency lanes with no usage limits.

Architecture

User → [Router] → [Daemon] → [SGLang] → [GPU]
          ↓
       [Redis]
          ↓
    [Go Backend]

• Backend: Customized PocketBase (auth, billing, daemon registry, scaling)
• Router: Rust edge server (auth, caching, unified API formatting, concurrency)
• Daemon: Rust SGLang wrapper (active request draining, health monitoring)
• Web: Astro + React dashboard for telemetry and account management