AI Tools

Track human-agent decisions with auditable approval chains

Open protocol that gives human-agent approvals, overrides and handoffs a shared, hash-chained audit trail across MCP and A2A.

gpt 4o mcpsqlitecursor

88 28 days ago

Test UI flows with autonomous agents that explore like users

Argus is a visual UI testing agent that explores a page like a real user through a six-stage pipeline, with no test scripts or selectors to maintain.

playwrightgeminisqlite

Review only the Python code you changed before pushing

Local Python code reviewer that checks only the files your next commit touches, no daemon or network.

git

claudegpt 4o openaianthropicgoogle

Build and extend code with a self-improving agent harness

Self-extensible coding agent harness with a unified multi-provider LLM API and a terminal UI library.

claude 3 5 sonnetclaude 3 opus google search consolevercelgithub

Go to market Tool

Build AI-powered SEO content engine with GSC feedback loops

Toolkit of 20 Node.js scripts and 14 JSON configs for an AI content engine driven by a weekly Google Search Console feedback loop and publish gates.

Content & media Tool

Generate 10-second 1080p videos with audio from text or images

Inference code for MAGI-2 Preview, a 114B-parameter MoE model that generates 10-second video with audio from text or an image, up to 1080p.

huggingfacedockerffmpeg

88 3 months ago

Research & summarize Tool

Research questions with verified citations and budget control

Deep research agent that verifies every claim against source text, enforces spending budgets to the cent, and keeps local data analysis private.

sqlitetavilybrave

88 2 months ago

Research & summarize Tool

Accelerate LLM inference in macOS VMs with Metal GPU unlock

Research release: a process-scoped Metal capability shim that speeds llama.cpp inference in macOS VMs on Apple Silicon by 7-16x.

gpt 4ollama 3 llama.cppmetalvirtualization.framework

claude slackdiscordgithub

Run autonomous coding agents in your terminal or CI pipeline

Self-contained Rust coding agent for your terminal: one small binary, any model or provider, fully offline with local GGUF models.

Track line-level authorship in AI-edited code

us-vs-them scores each line of a text 0-1 for human vs. agent authorship, derived from git history, to flag what agents should not overwrite.

git

91 2 months ago

claudegpt 4o claudecursorkiro

Organize and resume AI coding sessions across multiple agents

Terminal session manager that indexes, tags, and resumes every Claude Code, Cursor, Kiro, and Codex session.

91 3 months ago

Fine-tune and post-train LLMs with one command

One-command LLM fine-tuning CLI that trains an 8B model on a 4 GB laptop GPU via layer streaming.

discordtelegramgithub