openlit/openlit 2,814
OpenDCAI/DataFlex 3,237 +12 today
DataFlex: a DeepSpeed cap justified by a README pin, and results measured on a benchmark subset
open-compress/claw-compactor 2,004
claw-compactor's weighted average is a plain mean, its diagram names six of fourteen stages, and its branch has been idle since April
numtide/llm-agents.nix 2,181 +2 today
llm-agents.nix: a Nix flake that packages AI coding agents for reproducible installs
NovaSky-AI/SkyRL 2,450 +1 today
SkyRL: a full-stack RL library split into four installable layers
msoedov/agentic_security 2,028 +2 today
agentic_security: a prompt-dataset scanner you drive from a TOML file
mll-lab-nu/RAGEN 2,826 +1 today
RAGEN and StarPO: A Multi-Turn RL Framework With Built-In Diagnostics for Reasoning Collapse
mixelpixx/KiCAD-MCP-Server 2,944 +15 today
KiCAD MCP Server: Driving KiCAD 9 from an LLM Client
mgechev/skills-best-practices 2,278
skills-best-practices: a style guide for agent skills, not a runtime
meta-llama/PurpleLlama 4,418
Purple Llama splits LLM safety into moderation models and attack benchmarks
McGill-NLP/llm2vec 1,716
LLM2Vec: Turning Decoder-Only LLMs into Text Encoders
mbzuai-oryx/Video-ChatGPT 1,512
Video-ChatGPT: a video conversation model and its evaluation benchmark
lumina-ai-inc/chunkr 4,154
Chunkr: Self-Hosted Document Parsing for RAG Pipelines
Luce-Org/lucebox 2,953
Lucebox: a speculative inference server for consumer GPUs and mixed AMD/NVIDIA machines
llm-as-a-verifier/llm-as-a-verifier 3,334
LLM-as-a-Verifier: pairwise scoring for agent trajectories, and where it stops being free
latitude-dev/latitude-llm 4,734
Latitude: Trace-Level Observability That Hands Failures to Your Coding Agent
langfengQ/verl-agent 2,399
verl-agent: Step-Independent Rollouts for Long-Horizon Agent RL
kyutai-labs/unmute 1,521
Unmute: Kyutai's Websocket Bridge Between Text LLMs and Real-Time Speech
KouriChat/KouriChat 3,297 +2 today
KouriChat 1.5: an LLM companion bot that talks through OneBot, not Tencent apps
karthink/gptel 3,552