Odysseus NEW

Odysseus is a self-hosted AI workspace for running and serving local LLMs, autonomous agents, and multi-turn chat.Local-first, privacy-first architecture lets developers, researchers, and teams run models on private hardware or connect to external endpoints while retaining data control.

Model management includes a 270+ model catalog, hardware-aware recommendations, and one-click serving with MCP-compatible model serving across machines.Built-in tools (bash, files, web, memory) and toggleable tool permissions enable agents to plan, call tools, and complete multi-step tasks.

Productivity features include an email assistant (summaries, style-matched drafts, auto-tagging), document editor, notes & tasks, image gallery with inpainting, and scheduled agents.Research workflows support multi-step source gathering, reading, synthesis, and cited report generation; side-by-side compare evaluates multiple model responses.

Persistent memory, skill authoring, IMAP/SMTP support, optional telemetry, and external integrations enable long-term context, automation, and integration into existing infrastructure.

Key Features

• Self-hosted workspace for running and serving local LLMs, autonomous agents, and multi-turn chat
• Local-first, privacy-first architecture supporting private hardware deployment and external model endpoints
• Model management with model catalog, hardware-aware recommendations, and one-click MCP-compatible multi-machine serving
• Built-in agent tools (bash, files, web, memory) with toggleable tool permissions for multi-step task execution
• Persistent memory, skill authoring, IMAP/SMTP support, external integrations, and optional telemetry for long-term context and automation

Use Cases

• Host and serve private, multi-turn chat assistants and autonomous agents on-premises using Odysseus, preserving data privacy and compliance while leveraging hardware-aware multi-machine model serving to scale performance and reduce inference costs
• Build reproducible research workflows and generate cited, source-backed reports using Odysseus's model management, persistent memory, and built-in tools for evaluation, versioning, and exportable documentation
• Deploy internal AI tools—knowledge bases, customer-support agents, and analyst assistants—using Odysseus's integrations, persistent context and local LLM hosting to provide accurate, up-to-date answers without sending sensitive data to third-party services

Who is it for?

• Developers
• Machine learning engineers
• Data scientists
• Devops/system administrators
• Privacy-conscious organizations