Agentic RAG for local and self-hosted document search: hybrid retrieval, reranking and multimodal RAG on embedded LanceDB, with Docling parsing and an MCP server
-
Updated
Nov 19, 2026 - Python
Agentic RAG for local and self-hosted document search: hybrid retrieval, reranking and multimodal RAG on embedded LanceDB, with Docling parsing and an MCP server
General-purpose AI agent that acts via code actions through a unified execution interface.
Unified execution environment for Python code, shell commands, and programmatic MCP tool calls.
⚡ Lightning-fast file search MCP server using NTFS Master File Table - WizFile performance for Claude Desktop
Code Mode for MCP, where the sub-model is a calibrated decision model (Jev), not an LLM. Benchmarked on MCPToolBench++, LiveMCPBench, When2Call and live Claude agents.
A benchmark of what Claude Code's experimental Code Mode changes: six tests run with Code Mode off and on across four models, every number read from saved transcripts.
Python SDK and LangChain / Deep Agents middleware for Submilli: code mode for agents, with TypeScript programs that run on submilli-server under a Blueprint's policy.
🧩 Generate TypeScript code from any MCP server, with AST-Analysed Code Control & Security
Notion Knowledge Research. Use Notion with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Shopify Orders Report. Use Shopify with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Linear Backlog Review. Use Linear with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
HubSpot Pipeline Review. Use HubSpot with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Morning Brief. Use Gmail and Google Calendar with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Gmail, Google Calendar, Google Drive and Google Sheets MCP setup recipes for Google ADK (Python). Connect a remote Weavz MCP server with the native framework adapter, Tool Mode or Code Mode, scoped auth, and local fixture tests.
Support Knowledge Assistant. Use Gmail and Notion with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Salesforce, Jira Cloud, Google Drive and Slack MCP setup recipes for Semantic Kernel. Connect a remote Weavz MCP server with the native framework adapter, Tool Mode or Code Mode, scoped auth, and local fixture tests.
Salesforce Account Brief. Use Salesforce with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Todoist Day Plan. Use Todoist with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
Intercom Conversation Brief. Use Intercom with PydanticAI and Weavz MCP for bounded app reads, typed reports with source references, and synthetic local tests.
To associate your repository with the code-mode topic, visit your repo's landing page and select "manage topics."