Search

Local-first pentest TUI agent. Ink + Ollama. Operator-grade terminal for authorized security testing.

A CLI tool that traces git blame → PRs → issues → reviews and explains legacy code decisions in plain English using AI

Hardware inspector for Avasis Agent Builder. Detects CPU, RAM, GPU and recommends small models that run on your device.

Type-safe TanStack DB collections and TanStack Query options for Supabase tables, views, and RPC functions

React bindings for FactoredUI — provider, hooks, path context, SDUI renderer

AI Code Security Auditor — catches vulnerabilities that LLMs introduce and SonarQube misses. Purpose-built for AI-generated code with educational feedback.

UseSteady — safe refactor engine with step-by-step approval and full audit

AI-assisted development loop engine CLI

AuraFlow CSS — 49 drop-in design-system themes as pure CSS custom properties. Zero runtime, framework-agnostic. (Not the AI image model.)

daVinci-MagiHuman AI video generator. Create lifelike AI videos with daVinci-MagiHuman. Turn prompts or images into human-centered clips with smooth motion and fast online generation.

A pure UI dialogue box component for AI chat applications, built with React and Tailwind CSS.

Evaluating and steering alignment depth in LLM pre-training

Responsive, theme-based style props for building design systems with React

Official Chat SDK for the Persona API — React runtime provider with REST, WebSocket and WebRTC protocols via assistant-ui.

WorqApp Web chat SDK for it's own web

Redocly CLI OpenTelemetry

Offline context processing for LLM message arrays — compress, guard, meter

High-performance C++ tokenizer and prompt compressor for LLMs — reduce API costs by 10-25%

Convex component for server-side analytics tracking with Convalytics

Light weight browser analytics client intended to be used with the formation-web-analytics server.