unturf.

free software so you may grow food not lawn

uncloseai.

Free LLM & Text-to-Speech machine learning service powered by Hermes 3 (general purpose conversational automated intelligence) & Qwen 3.6 27B (vision & text). A drop-in replacement for OpenAI's API that requires no API key, making machine learning accessible to everyone. Supports both server-side & browser-based (uncloseai.js) implementations.

Mission: Machine learning for all, internet for all, open web for everyone.

Anonymous remote code, compile, & execution API for humans & machine learning agents. Write, generate, compile & execute code across 42+ programming languages in isolated containers. Returns code artifacts like binaries, images & more.

undefect.

Systematic defect research across open source. Catalogs structural defects replicated at scale, publishes patches upstream to fix them. Each MOAD (Mother of All Defects) names a class of defect found across hundreds of projects. All research, patches, & intel enter public domain.

Mission: Patch our planet. Every open source project scanned, patched, & disclosed makes shared infrastructure faster & cheaper for everyone.

Aggregates agent session logs & todo trackers for harnesses like Claude Code or uncloseai-cli & provides a dashboard to triage work across projects. Tracks machine learning usage & costs across projects & infrastructure spend across cloud & local inference offerings. A single pane of glass for agent analytics.

Web archival tool for digital preservation. Crawls domains, deduplicates content, captures screenshots, converts to markdown & creates self-extracting archives. Full-text search with intelligent metadata indexing & uncloseai.js integration.

arborist

An arborist for trees & forests of cross-linked information. Ingests documents into a content-addressed, Merkle-committed SQLite store, distills them into recursive cores, & answers questions over a corpus via an OpenAI-compatible LLM. Every cached answer carries a verifiable Merkle proof tying it back to source documents.

Mission: Merkle Providence Reverse RAG. Verifiable machine learning grounded in source.

A Lisp/Scheme-derived language centered on feedback as a primitive. Four independently implemented execution backends — Python, C, C+x86_64 JIT, and pure x86_64 GNU asm — run shared .lsp source byte-identically. A WebAssembly backend also supports Python, C, and asm targets. Features include Portals (continuations across processes), bend (GPU dispatch via CUDA), EML Universality (symbolic math), and MOAD Defect Isolation.