Odysseus
Self-managed AI workbench integrates conversation, Agent, document, memory, web research, mail, calendar, notes, tasks and local model tools.
Lazy cat OIDC login has been integrated. Odysseus is a self-hosted AI workbench suitable for use as a personal AI assistant portal in Lazy Cat Micro Clothing. It gathers capabilities such as chat, Agent, web research, model comparison, knowledge memory, notes, tasks, calendar, email and library in one interface, and saves data in its own micro-service environment. ##Who is suitable for use - Users who want to connect OpenAI, OpenRouter, Ollama, vLLM, llama.cpp and other models to one chat workbench. - Users who want to save chats, notes, tasks, knowledge bases and research materials on their devices. - Users who want to use Agent mode for multi-step tasks, web page retrieval, document organization, and content generation. - Users who want to compare the quality of responses from different models, or manage local/remote model services. ##Opening it for the first time 1. Install and start Odysseus in the lazy cat mini suit. 2. Open the app from the Lazy Cat app portal. 3. Log in using your lazy cat account. The application has been connected to the Lazy Cat OIDC. Under normal circumstances, the Lazy Cat login entry will be displayed or the application will be entered directly. 4. After entering the home page, if you see `Type /setup to get started.`, Note that no available models have been configured. 5. Configure the model first before starting chatting or using the Agent. ##Configuring AI Model Odysseus itself does not have a built-in large model and needs to be connected to external model services or local model services. Common access methods: - OpenAI API - OpenRouter - Ollama - vLLM - llama.cpp server - Other OpenAI API compatible services Recommendation process: 1. Click the lower left corner gear icon to enter settings. 2. Find the settings related to the model, endpoint, or Provider. 3. Add an OpenAI-compatible interface address and API Key. 4. Save and return to the chat page. 5. Select the model in the Select model in the upper right corner of the input box. 6. Enter the message and click the Send button. Example: ```text OpenAI: Endpoint: https://api.openai.com/v1 API Key: sk-... OpenRouter: Endpoint: https://openrouter.ai/api/v1 API Key: sk-or-... ``` If the model drop-down box is empty, there is usually a problem with the endpoint address, API Key, or network connection. ##How to use the sidebar function ### Search Used to search historical conversations, content or in-app data. Suitable for quickly retrieving previous questions, answers and information. ### Chats Manage all chat sessions. You can create, switch, and review historical conversations. ### Email After connecting to an IMAP/SMTP mailbox, you can view messages, organize your inbox, generate draft responses, and summarize messages. You need to configure your email account in your settings before using it for the first time. ### Tools Manage in-application tools and capabilities available to agents. When you need to perform complex tasks, you can first check whether the relevant tools are enabled. ### Brain Used to manage memory and knowledge. Odysseus will combine ChromaDB to provide vector memory capabilities suitable for storing long-term preferences, project background and reusable information. ### Calendar Used to view and manage schedules. After configuring CalDAV, you can synchronize with calendar services that support CalDAV. ### Compare Used to compare answers from multiple models. Suitable for testing the performance of different models on the same problem, helping to select models that are more suitable for writing, coding, summarizing or reasoning. ### Cookbook Used for model recommendation, download and service management. Suitable for users with a local GPU or a remote inference server. When using it for the first time, you need to configure it according to your own hardware and model service environment. ### Deep Research Used for multi-step data retrieval and research report generation. Suitable for investigating products, technical solutions, competitive products, market information or long problems. Make sure that web search capabilities are available before using. ### Gallery Used to manage pictures and generate results. Suitable for saving visual materials, screenshots or AI-generated pictures. ### Library Used to organize documents and materials. It can be used as a personal database to work with chat and Agents to summarize, rewrite and retrieve. ### Notes Used to take quick notes. Suitable for storing ideas, content to be organized and temporary data. ### Tasks Used to manage to-do items and tasks. You can cooperate with the Agent to organize the plans in the chat into a task list. ### Theme Used to adjust the interface theme and display style.




