Projects & Libraries

A type-safe web search library across 7 engines, with concurrent multi-engine search, automatic fallback and caching.

Generates 23 statistical analyses and 50 visualizations automatically from any data source, with HTML reports in 6 languages.

Converts a wide range of document formats into structured text an AI can actually use — 5 OCR engines, smart chunking, 80+ formats.

An agent-first browser automation toolkit written in Rust — native CDP performance with Python bindings.

A Python library that runs a 21-stage agent pipeline directly on raw provider APIs. Its dual-abstraction architecture normalizes 5 LLM providers onto one message and tool spec, so every stage is directly controllable and fully introspectable — without LangChain or LangGraph.

ⓘ Versions from PyPI

The same 21-stage interface as geny-executor with the data layer reimplemented in Rust (PyO3). Swapping it in is a one-line import change; wheels ship for Python 3.10–3.14 on Linux, Windows and macOS.

ⓘ PyPI release pending

A React graph renderer built on Three.js and d3-force-3d: 2 GPU draw calls, Web Worker layout, custom GLSL shaders, a 360° keyboard camera and an analytics module — production-grade.

A Python-native headless browser engine that lets AI agents treat the web as a structured semantic graph rather than pixels. It embeds V8 (mini-racer) and reduces a page to an actionable state machine through a three-method interface — navigate, snapshot, act — plus 13 tools including a fetch/network data plane.