WindsurfAPI

WindsurfAPI, also called DevinAPI, connects coding assistants and other AI clients to models through your own Devin account. The gateway runs on your hardware, but it sends prompts and images to Devin's cloud for inference, so it requires internet access.

• Client compatibility: Use OpenAI Chat Completions, OpenAI Responses, Anthropic Messages or Gemini clients, including Claude Code, Cursor and Cline.
• Model choice: Access Claude, GPT, Gemini, DeepSeek, Kimi, GLM and SWE models through one service, with a model catalog that reflects upstream availability.
• Images and tool calls: Send image inputs to supported models and pass tool calls between the model and your coding assistant; the client executes file operations locally.
• Account pooling: Distribute requests across accounts you control, with rate limiting, health checks and failover when an account becomes unavailable.
• Web dashboard: Manage accounts and model access, inspect live logs, and review request statistics by model or account.

The open-source gateway uses the MIT license and runs on Windows, macOS or Linux with Node.js, or in Docker. A packaged Windows app is also available. The default connection to Devin needs no separate language-server binary; the optional Windsurf language-server backend is Linux-only. You supply your own upstream account and credentials. Model access depends on that account, and the software license doesn't include access to Devin's services.