Run frontier LLMs directly inside Cursor, VS Code, Windsurf, or Claude Desktop using free browser login sessions or local offline providers. Ships with a self-healing Python agent, multi-agent delegation, and cross-session memory.
MCP.Pizza Chef: Zen4-bit
Proxima lets Cursor, VS Code, Windsurf, or Claude Desktop tap into ChatGPT, Claude, Gemini, and Perplexity by reusing the browser accounts you are already signed into, so you get access to several AI models without stacking separate paid subscriptions. It also works with fully local, offline AI models like Ollama, so your work never has to leave your computer. Setting it up means installing a small program on your computer and adding it to your coding tool's settings, so it takes a bit of technical comfort to get running.
Coding
Use This MCP server To
Chat with ChatGPT, Gemini, or Perplexity inside my coding tool Compare answers from different AI models before choosing one Skip paying for multiple AI subscriptions to get coding help Keep my AI assistant's memory across different coding sessions Run a private, fully offline AI model on my computer Hand off a multi-step coding task to a self-fixing AI agent
README
Proxima FAQ
Which apps does Proxima work with?
Proxima connects to Cursor, VS Code, Windsurf, and Claude Desktop, letting each one call out to ChatGPT, Gemini, or Perplexity in addition to Claude.
Do I need to pay for separate AI subscriptions to use this?
No — Proxima can reuse the free, logged-in browser accounts you already have for ChatGPT, Gemini, Claude, and Perplexity, so you do not need to buy extra plans just to use it here.
Do I need a paid account or a special access code to get started?
Not necessarily. Proxima's default mode reuses your existing logged-in accounts, though it also offers an optional mode where you can add your own paid access instead if you would rather use that.
Can I use this to get a second opinion from a different AI model without leaving my coding tool?
Yes — you can ask a question and route it to ChatGPT, Gemini, Claude, or Perplexity from inside the same coding tool, so you can compare answers without switching windows.
Can I use this to keep my work fully private and offline?
Yes — Proxima can run fully local, offline AI models such as Ollama, so nothing about your project has to be sent to any outside company.
How hard is this to set up?
It takes some technical comfort. You install a small program on your computer using a common developer install method, then point your coding tool at it, so it suits someone who is already comfortable coding.
Will this work on my computer?
Yes — Proxima runs on Windows, macOS, and Linux.
Does it store my passwords?
No — Proxima reuses your already-open browser sessions rather than saving your passwords or login details.