Documentation
The Proxyma user guide and integration documentation are included with the application and available inside the Proxyma web UI after installation.
In the meantime, if you have questions about setup, configuration, or MCP integration, reach out directly.
Getting Started
- Installation (Windows)Download, run the installer, and launch Proxyma Desktop
- Installation (Linux)Extract the .tar.gz archive and run the startup script
- First launch & Ollama setupConnect your local AI model on first run
- Activating your licenseInstall a license file to unlock Personal or Business features
Connecting Sources
- Local foldersRegister folders, file type filters, and recursive scan options
- Confluence spacesConnect with a personal access token and configure sync intervals
- Jira projectsIndex issues and attachments from Jira Server or Data Center
- Bitbucket, SharePoint, Teams, SVNBusiness tier connectors
MCP Integration
- GitHub Copilot in VS CodeAdd the MCP endpoint to your copilot-mcp.json config
- Claude CodeRegister Proxyma as an MCP server in .claude/settings.json
- Cursor & WindsurfConnect via Streamable HTTP MCP in IDE settings
- MCP endpoint referencehttp://localhost:4246/mcp — tools: search, getStatus, readTextFile
Search & Chat
- Hybrid search explainedBM25 + semantic vector search with RRF fusion
- Query expansion & rerankingAI-powered precision boost — enable in Preferences
- Built-in chat panelReAct agent with citations, multi-turn memory, @mention
- Supported AI providersOllama, Anthropic, OpenAI, Azure, DeepSeek, Mistral, Gemini...
Full documentation is available inside the Proxyma web UI at http://localhost:4246 after installation.