Stars
Forks
Watchers
Developer links
Odysseus
Agents with tool use, deep research, a document editor, an IMAP/SMTP email client with AI triage, notes, tasks, and a CalDAV-synced calendar - Odysseus bundles all of it into one open-source, self-hosted AI workspace. It runs local models through Ollama, vLLM, or llama.cpp and cloud APIs like OpenAI and OpenRouter, with a hardware-aware Cookbook that scans your machine and recommends quantized models that fit. Persistent memory uses ChromaDB with hybrid vector-plus-keyword retrieval, web search runs through a bundled SearXNG instance, and agents can use MCP servers, files, and shell access with safety controls, plus custom skills and scheduled agent tasks. A blind Compare mode runs side-by-side model duels with identities hidden and accumulates Elo-style ratings from your votes, so model selection is based on your actual workloads rather than leaderboard claims. Deep research mode - adapted from the Tongyi DeepResearch approach - reads sources through SearXNG and produces cited reports, while the email client tags, summarizes, sets reminders, and drafts replies locally rather than through a third-party mail AI. The writing-first document editor adds AI edits, Markdown and HTML support, and version history. The stack is Python 3.11 with FastAPI, SQLite for state, and a vanilla JS frontend, licensed AGPL-3.0 with zero telemetry. Because agents can read email and execute commands, keep authentication enabled and never expose it as a public unauthenticated service.
Benefits
- One Workspace Instead of Five Tabs
- Chat, research, writing, email, notes, and calendar live in a single self-hosted app, so the assistant works across the actual parts of your day rather than being an isolated chatbot.
- Zero Telemetry, Full Data Control
- Conversations, documents, email contents, and vector memory stay on your instance. With local models, nothing about your work leaves the server at all.
- Evidence-Based Model Selection
- Compare mode hides model identities during side-by-side tests and tracks Elo-style ratings from your votes, so you pick models based on your workloads instead of leaderboard claims.
- Memory That Persists
- ChromaDB-backed memory stores conversations, documents, and facts across sessions with hybrid retrieval, so context accumulates instead of resetting every chat.
Features
- Agents with Real Tools
- Agents plan and iterate using MCP servers, bundled SearXNG web search, file operations, and shell execution with safety controls, plus custom skills and scheduled agent tasks.
- Hardware-Aware Cookbook
- Scans your hardware, recommends compatible quantized models, and handles downloading and serving through Ollama, vLLM, or llama.cpp.
- Deep Research Mode
- Multi-step web research that reads sources through SearXNG and produces cited reports, adapted from the Tongyi DeepResearch approach.
- Email with AI Triage
- A real IMAP/SMTP inbox with tagging, summaries, reminders, and drafted replies, processed locally rather than through a third-party mail AI.
- Writing-First Document Editor
- Tabbed editor with AI edits and suggestions, Markdown and HTML support, CSV handling, syntax highlighting, and version history.
- Blind Model Comparison
- Run two models on the same prompt with identities hidden, vote on results, and accumulate per-model Elo ratings over time.