Familiars Documentation
Welcome to the Familiars documentation. Familiars is a Windows-first multi-agent AI desktop application that bundles dialogue-tree memory, scope×category long-term memory, browser/multimodal/voice tools, local QLoRA models, digital-human rendering with GPT-SoVITS, ASR (faster-whisper), social-platform listeners (WeChat / QQ / Feishu), and scheduled report generation.
Getting Started
New to Familiars? Start here.
| Page | What you'll learn |
|---|---|
| Installation Guide | Set up Python, Node.js, GPU drivers, and install all dependencies |
| Quick Start (5 min) | Launch the app, create your first agent, send your first message, create a Todo, use the CLI |
| Interface Tour | Walk through every major UI component in the desktop shell |
| FAQ | Answers to 40 common questions about installation, usage, models, network, privacy, and GPU |
Architecture
A ten-part deep dive into the technical internals, based on the repository's ARCHITECTURE.md.
- Architecture Overview — System topology, frontend, bridge, agent pipeline, RAG, storage, decider, digital human, security, ecosystem, and engineering appendix
User Manual
Step-by-step operational guides for daily use.
- Manual Index — Full user manual covering every feature
Developer Reference
For developers extending or integrating with Familiars.
| Page | What it covers |
|---|---|
| Bridge REST API Reference | Complete endpoint reference for the FastAPI bridge (60+ route files, 500+ endpoints) |
| CLI Command Reference | main.py command-line interface — 40 command groups, 200+ subcommands |
| Development Setup | Local development environment guide |
Ecosystem
Build recipes, flows, and broker apps on top of the Familiars platform.
| Page | What it covers |
|---|---|
| Recipe Development Guide | Build single-skill automation graphs (state-machine DAGs with LLM nodes, tool calls, variables) |
| Broker App Development Guide | Create shareable .broker.zip application packages with manifest + UI + plugins |
| Flow Development Guide | Cross-agent multi-stage workflows with parallel / branch / sub-track support |
Reference
- Security Model — Cloud data flow, local privacy, credential storage
Language / 语言
🌐 中文文档 | English