
CodeWhale
DeepSeek + MiMo coding agent in terminal
The Lens
CodeWhale is an AI coding agent that lives in your terminal. You talk to it, and it reads, writes, and runs code in your project, with approval gates so it can't run wild. The twist that sets it apart: it's built first around open models like DeepSeek V4 and Xiaomi's MiMo rather than Claude or GPT, which makes it the budget-conscious option in a category dominated by frontier-model tools. The agent itself is MIT licensed and free, and it's one of the fastest-moving projects in this space right now.
It's a serious piece of engineering, not a wrapper. Three modes (Plan is read-only, Agent gates every action, YOLO auto-approves), approval-gated access to files, shell, git, web, and MCP servers, side-git snapshots so you can roll back anything it changed without touching your real history, live diagnostics after edits, and parallel sub-agents. You install it with a single npm, Homebrew, or Docker command, no Rust toolchain needed. The one real setup step is getting an API key from DeepSeek, or another supported provider, and funding it.
That last point is the whole pricing story. The tool is free; the model usage is not. You bring your own key and pay the provider per token, and the entire pitch is that DeepSeek's pricing sits among the cheapest of the major providers, so the same agentic workflow costs a fraction of what it would on premium models. Solo developers who want a Claude Code style agent without the premium bill: this is the move. Small teams get visibility into what the agent changed before it lands. The closest comparisons are Claude Code, Aider, and OpenAI's Codex CLI, all of which lean on pricier models.
The catch is that you're betting on a young, blazing-fast-moving project tied closely to the open-model ecosystem. It ships releases almost daily, which means new features and occasional churn, and its best economics assume DeepSeek's API stays cheap and available. If you want the absolute strongest coding model regardless of cost, the frontier tools still win. If you want most of that quality for a fraction of the price, this is the bet.
Free vs Self-Hosted vs Paid
free self hosted paid cloudFree (the tool): MIT licensed and fully open source. Install via npm, Homebrew, Cargo, or Docker. No paywall, no hosted tier, no seat fees.
Paid (the models): Bring your own API key. You pay the model provider (DeepSeek, OpenRouter, NVIDIA NIM, Xiaomi MiMo, and others) per token. DeepSeek-first is the point: its per-token pricing is among the lowest of the major providers, so an agent that burns a lot of tokens stays affordable.
The trade: The same terminal-agent workflow as Claude Code or Aider, routed to cheaper open models. You give up frontier-model quality for a large cut in cost.
The agent is free and open source. You pay only for model usage, and routing it to DeepSeek keeps that bill low.
Get tools like this every Wednesday
One featured tool, three on the radar. No fluff.
Similar Tools

an open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM

The open source coding agent.

Autonomous coding agent right in your IDE, capable of creating/editing files, executing commands, using the browser, and more with your permission every step of the way.

aider is AI pair programming in your terminal
License: MIT License
Use freely, including commercial. Just keep the license.
Commercial use: ✓ Yes
About
- Owner
- Hunter Bown (User)
- Stars
- 37,952
- Forks
- 3,267
Explore Further
More tools in the directory
everything-claude-code
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
212.9k ★hermes-agent
The agent that grows with you
190.3k ★ollama
Get up and running with Kimi-K2.5, GLM-5, MiniMax, DeepSeek, gpt-oss, Qwen, Gemma and other models.
173.7k ★
