A 0.8B end-to-end document parsing model that generates Markdown from images.

Compile Lua to Native Executables Through C++20

AI-powered agents that automatically repair failing Playwright browser automation scripts.

Open-source observability and analytics for LLM applications.

Tool for testing and evaluating prompt quality and LLM outputs.

A tool for running open-source LLMs locally with an agentic framework

A text-UI disk usage treemap visualizer.

A collaborative multi-agent framework for the systematic assessment of controllability in recommender systems.

A dynamic deepfake detection system trained through adversarial competition to maintain robustness.

A tool for identifying semantic-preserving commits through behavior-based analysis and LLM-generated tests.

An extensible guardrail framework for securing agentic AI against prompt injection and tool misuse.

An open-source, lightweight infrastructure for evaluating LLM-based agents.

A curated list of resources and research on autoformalization of natural language to formal languages.

An R package for solving permutation puzzles using cycle intersection in Cayley graphs.

A record- and token-level data provenance system for AI training datasets to facilitate precise unlearning.

Run Arm64/x86-64 Linux ELF binaries on macOS Apple Silicon

A simulated IBM PC-compatible environment implemented entirely within a CSS stylesheet.

Source code and dataset for synthesizing execution-grounded trajectories for OS agents.

A benchmark task and artifacts for evaluating AI-assistability in multi-agent frameworks.

A framework to synthesize multi-tabular relational databases from scratch using directed and bipartite graphs.