Ultra-minimal AI chat UI: 30s deploy, macOS version; no sign-up; OpenAI-compatible; plugins/adapters.

The first open-source project to use GitHub Release as a transient middleware: Turn iOS captures into structured knowledge with zero-server cost.

Automated Triton w8a8 block FP8 kernel tuning tool for vLLM. Auto-detects model architecture, supports Qwen3-Coder-30B-A3B-Instruct-FP8/DeepSeek-V3/custom models, multi-GPU parallel tuning, and generates optimized kernel configs for quantization.

[fixed]vLLM MoE (Mixture of Experts) model kernel performance optimization tool

A Dify plugin for RMinte inference engine. Provides LLM, embedding, and reranking models with zero-configuration deployment.

Automated vLLM server parameter tuning tool. Finds optimal max-num-seqs and max-num-batched-tokens to maximize throughput. Includes presets for Llama/Qwen/Mixtral, batch processing, result analysis with visualizations, and environment checks. Supports TPU/GPU with latency constraints.