Chat application with multi-agents system supports multi-models and MCP
# Add to your Claude Code skills
git clone https://github.com/saigontechnology/AgentCrewAgentCrew lets you build and run a team of specialized AI agents instead of relying on a single generic assistant.
macOS / Linux
curl -LsSf https://agentcrew.dev/install.sh | bash
Windows
powershell -ExecutionPolicy ByPass -c "irm https://agentcrew.dev/install.ps1 | iex"
pip (any platform)
pip install agentcrew-ai
AgentCrew needs at least one AI provider key. Pick your preferred provider and add the key:
Option A — Environment variable
export ANTHROPIC_API_KEY="sk-ant-..."
# or
export OPENAI_API_KEY="sk-proj-..."
# or
export GEMINI_API_KEY="AIza..."
Option B — Config file
mkdir -p ~/.AgentCrew
cat > ~/.AgentCrew/config.json << 'EOF'
{
"api_keys": {
"ANTHROPIC_API_KEY": "sk-ant-..."
}
}
EOF
Option C — Subscription login
If you have a ChatGPT Plus / Pro subscription or a subscription, you can log in directly instead of using an API key:
No comments yet. Be the first to share your thoughts!
# ChatGPT Plus / Pro (Codex models)
agentcrew chatgpt-auth
agentcrew chat --provider openai_codex
# GitHub Copilot
agentcrew copilot-auth
agentcrew chat --provider github_copilot
Supported providers: Anthropic Claude, OpenAI, Google Gemini, GitHub Copilot, DeepInfra, Together AI, OpenCode, and any OpenAI-compatible endpoint.
Tip: Not sure which provider to pick? Claude and OpenAI work great for most users. See CONFIGURATION.md for provider-specific setup details.
# Desktop GUI
agentcrew chat
# Terminal mode
agentcrew chat --console
On the first launch, AgentCrew will walk you through creating your first agent if you do not already have one.
If you already have an API key set and want to create a new agent from scratch:
agentcrew create-agent
Or define one manually in ~/.AgentCrew/agents.toml:
[[agents]]
name = "CodeAssistant"
description = "Helps write and review code"
tools = ["code_analysis", "file_editing", "web_search", "memory"]
system_prompt = """You are an expert software engineer.
Focus on code quality, security, and maintainability.
Today is {current_date}."""
Switch between agents, attach files, and let your team handle the rest.
/agent Architect
Design a clean API for a task manager.
@Coding
Implement the task manager in Python using FastAPI.
@Reviewer
Review the code for security issues.
| Mode | Command | Best for |
| ----------------- | ----------------------------------------------------------- | ----------------------------------------------------------- |
| Desktop GUI | agentcrew chat | Daily interactive work, file drag-and-drop, visual diffs |
| Terminal | agentcrew chat --console | Remote servers, low-overhead use, keyboard-driven workflows |
| One-shot jobs | agentcrew job --agent "CodeAssistant" "your task" ./files | CI/CD scripts, automation, batch processing |
| HTTP API | agentcrew a2a-server | Integrating with other apps, multi-instance setups |
Job mode example:
agentcrew job --agent "CodeAssistant" \
"Review for security issues" \
./src/**/*.py
A2A server example:
agentcrew a2a-server --host 0.0.0.0 --port 41241
Agents come with a toolkit you enable per agent:
See the examples/agents/ folder for ready-to-use agent
setups. To use an example:
cp examples/agents/agents.simple.toml ~/.AgentCrew/agents.toml
agentcrew chat
AgentCrew stores settings in ~/.AgentCrew/:
| File | Purpose |
| ------------------ | ---------------------------------------- |
| config.json | API keys, theme, global preferences |
| agents.toml | Agent definitions, tools, system prompts |
| mcp_servers.json | External tool servers (optional) |
You can edit these files directly or manage them through the GUI settings panel.
Inside the chat interface, type:
/agent <name> — switch to another agent/clear — start a new conversation/file <path> — attach a file/copy — copy the last response/think <low|medium|high|xhigh> — enable reasoning mode/voice — start voice recording/help - show all available commandsexit or quit — close AgentCrewApache 2.0 License. See LICENSE for details.