Search2o 0.11.0 Adds a Self-Hosted Server for Enterprise AI Agents

Summary

Search2o 0.11.0 is an open-beta, stateless and asynchronous Python server for running task-specific AI agents within an organization. It can be installed with pip and includes a bundled GUI, a REST API, and OpenAPI, Swagger UI, and ReDoc endpoints. Users create an account to obtain a license key, start the server, and configure an account through a local web interface at 127.0.0.1:9020. Self-hosted models accessed through Ollama, vLLM, or SGLang do not require a vendor API key; OpenAI, Anthropic, and Gemini integrations require their respective environment variables. In the GUI, users select or create LLM profiles, draft agents with AI, validate and publish them, and add plain-English descriptions so the search interface can match and run an agent for a question. The same agents can also be used through the bundled search2o-skill in Claude Code. The package requires Python 3.12 or newer and offers optional extras for PostgreSQL, MySQL, Oracle, and Microsoft SQL Server. Search2o is source-available but proprietary, distributed under a proprietary software license, and the project currently does not accept external code contributions, although it welcomes bug reports, feature requests, documentation feedback, and security reports.