# Odysseus Self-hosted AI workspace for chat, autonomous agents, deep research, email, documents, and more — local-first, privacy-first, no telemetry. ## Agent Decision Summary - Risk level: moderate - Source confidence: high - Recommended workflows: Coding agent workflow, Local or private AI stack - Permission surface: memory, external services - Agent JSON: https://www.openagent.bot/agents/odysseus.agent.json ## Summary Odysseus is a self-hosted AI workspace that unifies chat, autonomous agents, deep research, email triage, document editing, calendar, notes, memory, and model serving into a single local-first interface. It runs on your own hardware against your own endpoints — no telemetry, no cloud dependency, and full data privacy. ## Guide Odysseus is a self-hosted AI workspace that brings together chat, autonomous agents, deep research, email management, document editing, calendar, notes, memory, and model serving into one local-first application. It runs entirely on your hardware against your own endpoints. ### What it is Odysseus is an open agent resource to evaluate by action surface: what software it can operate, which tools or browser steps it touches, and how much supervision it needs before it can run real work. ### Why it matters The self-hosted AI space has many single-purpose tools — a chat UI here, an agent framework there, a model server somewhere else. Odysseus combines them into a cohesive workspace, making it practical to run your entire AI stack locally without stitching together disparate services. ### How it works Start with one safe workflow for Odysseus. Inspect official setup instructions, required credentials, execution logs, approval points, and failure recovery before expanding from a sandbox task into production automation. ## Use Cases - Local AI workspace for daily workflows: Use Odysseus as your daily driver for chat, research, email, documents, and calendar — all connected to your own local or API-hosted models. - Autonomous task execution with MCP tools: Give the agent access to MCP servers, shell, files, and web search. Let it plan and execute multi-step tasks while you supervise. - Private email assistant: Connect IMAP/SMTP accounts for AI-powered summarization, auto-tagging, spam triage, and draft replies — all processed locally. - Model discovery and serving: Use the Cookbook to scan hardware, browse 270+ catalogued models by VRAM fit score, download, and serve them — all from the UI. ## Alternatives - Use Odysseus for breadth; dedicated tools for depth vs specialized agents: Aider, OpenHands, and Claude Code are more focused on coding. Odysseus offers a broader workspace — email, documents, calendar, model serving — alongside agent capabilities. - Use cloud workspaces for zero-setup convenience vs ChatGPT / Claude: ChatGPT and Claude require no setup and offer polished hosted experiences. Odysseus requires Docker or Python setup but keeps all data local with no telemetry. ### Getting Started - Read the README: https://github.com/pewdiepie-archdaemon/odysseus - Visit the official site: https://pewdiepie-archdaemon.github.io/odysseus/ ### FAQ - What should I check before using Odysseus? - Start with one safe workflow for Odysseus. Inspect official setup instructions, required credentials, execution logs, approval points, and failure recovery before expanding from a sandbox task into production automation. - Is Odysseus open source? - Odysseus is listed with MIT based on the official source links in this profile. Re-check the repository, model card, or docs before production use. - Who should evaluate Odysseus? - Odysseus is most worth evaluating for developers who want a self-hosted all-in-one AI workspace with full data privacy. - Does Odysseus require a GPU? - No. Odysseus connects to any LLM endpoint — local or remote. The Cookbook model serving is optional and GPU-dependent. You can use it with API providers like OpenAI, Anthropic, or OpenRouter without any local GPU. - How does Odysseus compare to running separate tools? - Odysseus combines chat, agents, deep research, email, documents, calendar, and model management into one application. The tradeoff is breadth over specialization — dedicated tools may go deeper on specific tasks. ## What It Does Odysseus is an open agent resource to evaluate by action surface: what software it can operate, which tools or browser steps it touches, and how much supervision it needs before it can run real work. ## How To Evaluate Start with one safe workflow for Odysseus. Inspect official setup instructions, required credentials, execution logs, approval points, and failure recovery before expanding from a sandbox task into production automation. ## Why It Matters Odysseus matters because it offers a rare combination: a full-featured AI workspace that runs entirely on your own hardware. Most AI tools push you toward a hosted subscription — Odysseus gives you chat, agents, research, email, documents, and model management in one self-contained package, all while keeping your data local. ## Best For - Developers who want a self-hosted all-in-one AI workspace with full data privacy - Users who run local models and need a polished UI for chat, agents, and tools - Teams evaluating self-hosted alternatives to cloud AI workspaces like ChatGPT or Claude ## Not For - Users who prefer managed cloud services with no self-hosting overhead - Non-technical users unwilling to manage Docker or Python runtimes ## What It Actually Does - All-in-one self-hosted workspace: Odysseus combines chat, autonomous agents, deep research, email triage, document editing, calendar, notes, memory, model serving, and more in a single application. - Why it matters: Most AI tools require stitching together multiple services. Odysseus gives you everything in one local-first package without sending data to the cloud. - Local-first with privacy by default: Runs entirely on your hardware against your own model endpoints. No telemetry, no cloud dependency, and no third-party access to your conversations or data. - Why it matters: Privacy is the default architecture, not an afterthought or a premium feature. - Agent loop built on opencode with MCP: Odysseus's agent system is built on opencode, with full MCP tool support, web browsing, file access, shell commands, and persistent memory. - Why it matters: The agent isn't just a chat wrapper — it can take real actions, use tools, and evolve its own skills over time. ## Typical Use Cases - All-in-one local AI workspace: Run chat, agents, email, documents, and calendar from a single self-hosted dashboard connected to your own models. - Autonomous agent workflows: Give the agent tools (MCP, shell, files, web) and let it plan and execute multi-step tasks autonomously. - Local model management and serving: Use the Cookbook to scan your hardware, discover suitable models, download them, and serve them — all from the UI. - Privacy-preserving email assistant: Connect your IMAP/SMTP accounts for AI-powered email triage, summaries, and draft replies — without sending email data to third-party services. ## How It Compares - Choose Odysseus for a complete self-hosted workspace vs cloud AI workspaces: ChatGPT and Claude offer polished hosted experiences. Odysseus gives you the same breadth of features — chat, agents, research, tools — but runs entirely on your own hardware with no telemetry. - Choose Odysseus for breadth over specialization vs single-purpose tools: Dedicated coding agents like Aider or OpenHands go deeper on code tasks. Odysseus is broader — it handles email, calendar, documents, and model management alongside agent capabilities. ## Fit Matrix - Coding agent workflow: strong. Odysseus has multiple signals for coding agent workflow, including matching tags, capabilities, category, or positioning. Required check: Run a small repository change and inspect the diff, tests, and rollback path. - Local or private AI stack: strong. Odysseus has multiple signals for local or private ai stack, including matching tags, capabilities, category, or positioning. Required check: Verify hardware requirements, data path, storage, and whether all calls stay in your environment. - Connector or protocol layer: partial. Odysseus has at least one signal for connector or protocol layer, but should be checked against a real task before adoption. Required check: Connect one low-risk service, then inspect schemas, auth scope, errors, and logs. - Evaluation and observability: partial. Odysseus has at least one signal for evaluation and observability, but should be checked against a real task before adoption. Required check: Add one repeatable test case and confirm results can run again in review or CI. - Memory or RAG workflow: partial. Odysseus has at least one signal for memory or rag workflow, but should be checked against a real task before adoption. Required check: Create, update, retrieve, correct, and delete memory or retrieval objects with real data. - Reusable skill workflow: partial. Odysseus has at least one signal for reusable skill workflow, but should be checked against a real task before adoption. Required check: Run one skill end to end and check whether it produces evidence or structured output. ## Evidence - verified: Odysseus is listed as open source. Source: License metadata: MIT - verified: Odysseus has a recorded GitHub repository: pewdiepie-archdaemon/odysseus. Source: Resource facts and GitHub source link. - inferred: Odysseus supports these recorded deployment modes: self hosted, cloud. Source: OpenAgent decision signal metadata. - inferred: Odysseus is tagged with mcp, workflow orchestration capabilities. Source: OpenAgent capability taxonomy. ## Missing Checks - Dedicated docs link is missing. - Repository freshness has not been recorded. ## Next Actions - Inspect repository: https://github.com/pewdiepie-archdaemon/odysseus - Open Homepage: https://pewdiepie-archdaemon.github.io/odysseus/ - Install with Docker (recommended): git clone https://github.com/pewdiepie-archdaemon/odysseus.git && cd odysseus && docker compose up -d --build - Install natively on Linux/macOS: git clone https://github.com/pewdiepie-archdaemon/odysseus.git && cd odysseus && python3 -m venv venv && source venv/bin/activate && pip install -r requirements.txt && python setup.py && python -m uvicorn app:app --host 127.0.0.1 --port 7000 ## Command Line ### Install with Docker (recommended) Clone the repository and start Odysseus with Docker. Open http://localhost:7000 when containers are healthy. ```bash git clone https://github.com/pewdiepie-archdaemon/odysseus.git && cd odysseus && docker compose up -d --build ``` ### Install natively on Linux/macOS Native install requires Python 3.11+. The Cookbook feature also needs tmux for background model operations. ```bash git clone https://github.com/pewdiepie-archdaemon/odysseus.git && cd odysseus && python3 -m venv venv && source venv/bin/activate && pip install -r requirements.txt && python setup.py && python -m uvicorn app:app --host 127.0.0.1 --port 7000 ``` ### Apple Silicon (GPU-accelerated) Docker on macOS cannot use Metal GPU. Use the native script for GPU-accelerated model serving on M-series Macs. ```bash git clone https://github.com/pewdiepie-archdaemon/odysseus.git && cd odysseus && ./start-macos.sh ``` ## Facts - Category: agents - Resource type: agent - Open source: yes - License: MIT - Last verified: 2026-06-03 - GitHub repo: pewdiepie-archdaemon/odysseus - GitHub stars: 36300 ## Capabilities - mcp - workflow-orchestration ## Structured Use Case Tags - self-hosted-ai - developer-workflow ## Getting Started - Open the GitHub repository: https://github.com/pewdiepie-archdaemon/odysseus - Visit the official site: https://pewdiepie-archdaemon.github.io/odysseus/ ## Links - GitHub: https://github.com/pewdiepie-archdaemon/odysseus - Homepage: https://pewdiepie-archdaemon.github.io/odysseus/ ## Structured Outputs - JSON: https://www.openagent.bot/agents/odysseus.json - Markdown: https://www.openagent.bot/agents/odysseus.md - Agent JSON: https://www.openagent.bot/agents/odysseus.agent.json - Canonical: https://www.openagent.bot/agents/odysseus