ethanhq

ethanhq / cc-fleet

Public

🚢 Spawn any vendor LLM — DeepSeek · GLM · Qwen · Kimi · MiniMax … — as real Claude Code teammates or ⚡ one-shot subagents 🚀 ｜🤖 让 Claude Code 接入任意第三方模型(DeepSeek · GLM · Qwen · Kimi · MiniMax ……),作为原生 Agent Team Teammate 或 ⚡ 一次性 subagent 为你干活 🚀

ai ai-agents anthropic anthropic-claude bash-cli
47
4
80% credibility
Found Jun 01, 2026 at 47 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Go
AI Summary

cc-fleet is an open-source tool that lets Claude Code work alongside third-party AI models (DeepSeek, GLM, Qwen, Kimi, and similar services) as intelligent teammates. It runs these AI assistants in separate tmux window panes, manages their API credentials securely, and teaches the main Claude Code session when and how to delegate work to them. Users can either spawn persistent 'teammate' AI workers that stay alive across conversations, or fire off quick one-shot requests to any provider. The tool handles all credential management, profile generation, and process coordination so users can focus on describing what they need rather than managing infrastructure.

How It Works

1
🔍 You discover a better way to use Claude Code

You hear about running multiple AI assistants at once — like having a whole team of experts working together on your project.

2
⚡ You install cc-fleet in one line

A single command downloads and sets everything up. Your computer is now ready to connect to other AI services.

3
🔌 You connect your favorite AI services

Through a friendly interface, you add DeepSeek, GLM, Qwen, or any AI provider. Each gets its own secure connection that never touches your command history.

4
🎛️ Everything configures automatically

The tool creates the right settings files in the right places. You never have to touch a configuration file yourself.

5
You choose how to work
👥
Teammate mode

Spawn a persistent AI teammate that lives in a window next to Claude. Hand it ongoing work and it keeps context.

⚡
One-shot mode

Send a quick task to any AI provider. Get your answer and done. No lingering sessions.

6
🚀 Claude coordinates everything for you

You speak to Claude Code naturally. Behind the scenes, it picks the right AI for each job, spawns teammates, collects results, and brings everything together.

✨ Your project gets done faster

Multiple AI perspectives working together, your work organized, your tools connected. Everything just works.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 47 to 47 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is cc-fleet?

cc-fleet is a Go-based CLI that lets you use any third-party LLM provider with an Anthropic-compatible API as a real Claude Code teammate. Instead of being locked into Claude's own models, you can spawn DeepSeek, GLM, Qwen, Kimi, or MiniMax workers that behave exactly like native teammates -- driven via TeamCreate and SendMessage. The tool manages vendor profiles, handles API key dispatch securely through a plugin system, and runs teammates inside tmux panes so you can watch them work alongside your lead session. It also supports one-shot subagent mode for headless batch tasks that don't need a full team session.

Why is it gaining traction?

The hook is simple: Claude Code is powerful, but you're stuck with its native models unless you bring your own API keys. cc-fleet breaks that lock-in without requiring you to abandon Claude Code's agent framework. You get the orchestration layer (teams, delegation, multi-turn memory) while picking any vendor's pricing and model capabilities. The interactive TUI makes vendor setup approachable, and the skill teaches Claude when to actually delegate work -- so it's not just a config file, it's a workflow. The security-first API key handling (keys never hit argv or shell history) addresses the trust concern that would otherwise make this a non-starter.

Who should use this?

Developers who want Claude Code's agent capabilities but prefer specific vendor models for cost, latency, or capability reasons. Teams evaluating multiple providers side-by-side will find the parallel teammate spawning useful. Power users who want to run batch analysis via subagents without spinning up full team sessions will appreciate the headless mode. If you're already happy with Claude's native models, this adds complexity without clear benefit.

Verdict

cc-fleet solves a real workflow gap for cost-conscious or vendor-preferring developers, and the implementation shows careful thought around security and concurrency. At 47 stars, it's early-stage and the documentation is functional rather than extensive -- test coverage is present but you'll want to read the code for edge cases. The 0.800000011920929% credibility score reflects a small but active project. Worth trying if you need multi-vendor Claude Code, but treat it as a power tool rather than a turnkey solution.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.