1,210 tools found
CLI that prices token waste in an AI agent trace, prescribes fixes, and can fail CI when a change makes the agent measurably more expensive.
cursor-bridge is a Rust binary that runs Claude Code CLI using your Cursor subscription backend through automatic proxy management.
AI-native OS running inference on your own hardware, federating peer-to-peer with no broker, exposing an OpenAI-compatible API on port 6777.
Staff-engineer control layer for AI coding agents: routes only the architectural constraints relevant to the file being edited, then requires verification.
claude-thermos keeps Claude Code's prompt cache warm while a subagent runs long, avoiding the expensive full re-encode when the main agent resumes.
libargus is a zero-allocation native AI runtime unifying LLM, Whisper ASR, TTS, and multimodal inference behind one Project Panama FFM Java boundary.
capn-hook is persistent memory for coding agents - it saves which files answer a question, so the next session skips re-exploring the codebase.
aict reimplements 33 Unix coreutils with structured XML/JSON output so AI agents never have to parse plaintext column formats.
Grepathy extracts undocumented decisions from Claude Code session transcripts into a committed markdown why-file, so PR reviewers can see the reasoning.
CLI that benchmarks local LLM inference, splitting prefill, decode, and wallclock speed, and cross-checks the OS GPU meter to catch silent CPU fallback.
AVA is an open-source AI voice agent for Asterisk/FreePBX with a modular STT/LLM/TTS pipeline and 7 golden baselines for cloud, local, or hybrid use.
AI Trains AI uses RL to train an agent that itself writes, submits, and runs RL training jobs for small models on real GPUs.
