Now in open beta — orchestrate a team of Mel agents from your terminal. Learn more

Mel Agent

Ask, and it does the work.

The agent that lives in your terminal. Describe a task; Mel plans, runs commands with your approval, reads files, and edits code — across 6 providers and 41 models, with no API keys to set up.

mel · agent
/agent add rate limiting to the orders API
auto → claude-sonnet-4-6 reading routes/orders.ts
✎ edit routes/orders.ts +18 −2
npm test ✓ 24 passed

What it does

A real engineer's loop, automated.

Ask in plain English

No flags to memorize. Describe the task and Mel figures out the commands, files, and steps.

6 providers, 41 models

Anthropic, OpenAI, Gemini, Groq, DeepSeek, Moonshot. Pick one, or let 'auto' route per task.

Cross-provider fallback

Hit a credit or rate limit? Mel falls back across providers automatically and keeps going.

You approve every change

Writes, edits, and commands round-trip an approval card — or flip on autonomous A⁺ mode.

Runs locally

Read, edit, and run all happen on your machine. Only the context a call needs is sent.

Verifies its work

Mel re-reads, builds, and tests until the change actually works — not just until it looks done.

Put the agent to work.

Free, open source, no API keys.