Reverse engineer anything with agents, from app behavior down to native binaries.

Codebase for Berkeley Humanoid Lite

Training Jev-like decision models using Qwen3 with choice-order invariance

Find code by asking what it does. A CLI for coding agents that uses Jev to discover relevant files and source context.

Isolated VM environment for running Claude Code and Codex

Release package for SlopShape: Identifying AI-Generated Commercial Web Content (verification artifacts, instrument, prompts, code)

Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per re…

Local computer use on Apple silicon: on-device typed decisions with laya and form filling with CUA-S1-FORMS, driven through the macOS Accessibility API

Chrome extension that filters X timelines with your own TypeSafe Jev API key, or GLiNER-Decide locally

Validator: a Rust CLI for evaluating classifiers and TypeSafe Jev workflows: single/multi-label metrics, error analysis, run comparison

Source code for the Claude Opus 5.5 music video for I'm Upping My P(doom)

SoL-Pi: Scaling Auto-Research Loops for Efficient Agent Harnesses

Static analysis tool that catches 1000+ bug patterns across all popular programming languages, with auto-wiring into AI coding agent quality guardrails

DuckDB extension for tabular foundation models — zero-shot classification & regression in SQL. Built-in Mitra, TabPFN v2, TabICL v2, and Google TabFM (selectable by name); register your own in SQL.…

TabFM (Tabular Foundation Model) is a pretrained tabular foundation model developed by Google Research for tabular data regression and classification.

A self-hosted drop-in replacement for TypeSafe's jev, powered by GliFormer.

Scale computer-use 2.0 with open-source drivers, cross-OS fleets, and benchmarks for training, evaluation, and data generation.

Browser Harness | Self-healing harness that enables LLMs to complete any task.

Real browser automation with local GLiNER2 inference and open weights. Work in progress. Contributions welcome.

Tax document page classifier built on Jev decisions. 100% strict accuracy across 261 IRS forms, ~$0.001 per page.

Automation foundation model for tiny devices: 2-bit, 8-29 MB, tool calls, structured extraction and embeddings on phones, wearables, smart homes, robots, cars and microcontrollers.

Generalist and Lightweight Model for Text Classification

Fully autonomous AI Agents system capable of performing complex penetration testing tasks

Removes 20+ patterns of AI slop from any piece of writing.

At last anyone can be a shape rotator! Visualize a codebase as an interactive dependency machine: jet-engine schematic, 3D orbit, circuit board, arc diagram, and staged pipeline views with animated…

LLM-as-a-Verifier is a general-purpose framework that provides fine-grained feedback for any agent without requiring additional training. It achieves SOTA performance across coding, robotics, and m…

Make your agents talk human. ADHD-friendly output styles for Claude Code, Codex, and others. So you can pay attention, not tokens.