DuckDB

Security-First Agentic IDE Coding Harness

A scope for your Ai coding-agent sessions

Deterministic, MCP-native charts for AI agents — AI-built, AI-maintained. 52 verbs, byte-stable SVG, DuckDB inside.

A bounded SQL-on-files CLI for coding agents. SQL in, files out.

AI-powered chatbot for exploring and querying OpenFoodFacts product data using natural language, SQL, and DuckDB.

An AI-powered Open Data Copilot that turns raw datasets into audit-grade decisions, automated EDA, cleaning workflows, SQL-backed insights, and vector PDF reports using DuckDB, Pandas, Groq LLM, and Google Drive BYOS.

AI Data Analyst agent (CSV/Excel/PDF/DOCX + live SQL DBs) and RAG document chat, built on a hand-rolled Gemini function-calling loop.

Runnable GeoLens integrations for MapLibre, QGIS, ArcGIS, Python, TypeScript, DuckDB, STAC, OGC APIs, CLI and MCP. Every live example runs against the public demo.

Plateforme d’analyse Airbnb intégrant dbt, DuckDB et Streamlit pour la transformation, le stockage et la visualisation interactive des données, avec une étude de l’impact de la pleine lune sur les avis clients.

Evidence-grounded AI agent for business analytics, forecasting, anomaly detection, and decision intelligence.

End-to-end analytics engineering project with dbt, DuckDB, LangChain AI agent and Streamlit.

“Job Market Radar: a data pipeline for job post collection + governed analytics API + Agent to analyze job postings.”

My personal Claude Code architecture — Skills, sub-agents, hooks, settings, MCP servers. Anchored on factforge.

Project for the Raiffeisen Group IT. A agentic system where you can connect your Postgres or DuckDB Database to ask quesitons about the data. Also includes a MCP Server that is connectable with Claude Desktop; copy from gitlab for showcase;

Conversational AI agent over accounts-receivable data: a LangGraph ReAct agent with guarded text-to-SQL and RAG, full-stack (FastAPI + React), on synthetic data

SwarmPIM is a production-grade, containerized PIM system engineered with Next.js 14 and FastAPI. It orchestrates a multi-agent AI swarm utilizing Gemini and Grok with dynamic fallback routing. The architecture features an embedded Qdrant vector cache, high-speed Polars batch ingestion, and analytical DuckDB storage optimized via Apache Arrow.

Lint the documentation & metadata quality of VGI (Vector Gateway Interface) data workers — descriptions, column comments, tags, and example queries — with a quality score, per-version baselines, and agent-friendly output.

Self-hosted AI data analyst — ask your CSV anything: agentic text-to-SQL over DuckDB with guardrails, streaming, and evals

Personal AI-news aggregator: RSS/HN to DuckDB to static site on GitHub Pages

End-to-end data engineering and analytics pipeline for Amsterdam Airbnb data using Python, DuckDB, SQL, and Streamlit.

Real-time support ticket pipeline using Redpanda, Quix-Streams, DuckDB, and an AI agent — ask questions about your streaming data in natural language.

Finance-owned FP&A tool where code computes every number and the LLM only explains — with a deterministic audit proving zero fabricated figures. Synthetic SaaS data.

Data pipeline + Analytics Dashboard for Historical Air Quality Index (AQI) data

End-to-end data pipeline monitoring air quality across 5 European countries using dlt, DuckDB, dbt & Streamlit

Por donde entran tus datos. Ingesta conversacional sobre DuckDB: la IA diseña la carga, el SQL la ejecuta sola. / Conversational data ingestion on DuckDB. The AI designs the load; SQL runs it. Local-first, no cloud. Local, sin nube, sin servidor.

Diagnóstico de datos de una sola pasada: sube tus archivos desordenados y obtén el modelo real de tus datos, un diagnóstico en lenguaje de negocio y los cruces posibles — todo con evidencia SQL verificable. DuckDB decide, el LLM narra.

Isolated PyAirbyte proof of concept for GitHub PR metrics in DuckDB and Databricks Delta tables

duckdb2-agent-smith (Ad-Free)

Use this skill when the user wants to turn a flat, messy clinical CSV into a medical OLAP data warehouse. It guides the agent to read the raw data, design Star, Snowflake and Fact Constellation schemas around a central Visit spine, load the data into DuckDB, and use sub-agents to prove that nothing was lost or changed.