AutoPilot AI v1.0.0

🚀 AutoPilot AI v1.0.0 — Initial Release

Password:AI

The first public release of AutoPilot AI — an open-source autonomous AI agent framework for Python.

---

✨ Highlights

• Minimal & Fast — ~2K lines of code, 2-minute setup
• Multi-LLM Support — OpenAI, Anthropic, Ollama, Groq via LiteLLM
• Multi-Agent Pipelines — chain Researcher → Coder → Writer → Analyst
• Long-Term Memory — ChromaDB-backed vector store for persistent context
• Plugin System — Gmail, Slack, GitHub, Notion integrations
• REST + WebSocket API — FastAPI server with real-time streaming
• Beautiful CLI — Rich terminal UI with ASCII banner, progress, panels
• Docker Ready — one-command deployment with docker-compose

---

🧠 Core

Component	Description
Engine	Plan → Execute → Synthesize loop with ReAct-style tool use
Planner	LLM-powered task decomposition into 1–5 steps
Memory	ChromaDB vector store with cosine similarity recall
Tool Registry	Decorator-based @tool system with auto-generated OpenAI schemas

🤖 Agents

Agent	Role
GeneralAgent	Ad-hoc tasks, default for Agent() shortcut
Researcher	Web search, browsing, source analysis
Coder	Code generation, execution, debugging
Writer	Documentation, articles, reports
Analyst	Data analysis, insights, recommendations
Pipeline	Sequential multi-agent orchestration with context passing

🔧 Built-in Tools

• web_search — SerpAPI / DuckDuckGo fallback
• code_executor — sandboxed Python execution (30s timeout)
• file_manager — read, write, append, list, delete
• browser — Playwright headless page extraction
• api_caller — HTTP GET/POST/PUT/DELETE

🔌 Plugins

• GitHub — create issues, list open issues
• Gmail — read, search, send (scaffold)
• Slack — post messages, read channels (scaffold)
• Notion — create/update pages (scaffold)

🌐 API

• POST /api/v1/run — execute a task, get structured result
• GET /api/v1/health — health check
• WS /ws — WebSocket for real-time streaming

🖥️ CLI Commands