Search

A local-only Ollama inference benchmark implementing the osai-bench measurement protocol.

Cross-platform npm entrypoint for installing coral-code via the official installer scripts.

本地真实 LLM Agent 通过 WebSocket 连接 Worker/MatchDO。CLI 不持有本地引擎， 撮合、座位、投影视图和动作提交全部走生产协议。

DeepInfra provider for pi: dynamic model catalog, reasoning-effort thinking levels, vision, session usage + monthly billing footer

Drop-in production trace collector for LLM agents (TypeScript/JS). Streams Vercel AI SDK, Anthropic, OpenAI, and LangChain.js traces to your Mirrors backend with negligible overhead.

Pisces — The CS Student Edition for Pi Coding Agent. Your personal AI teaching assistant from homework to thesis.

Schema/DSL-driven React map visualization component library powered by L7

Sentinel CLI — scan your codebase for API integration problems

Agent-first command line client for SealAI

Detect LLM-assisted 'cyborg' behavior in online behavioral experiments

Local CLI for reducing context and tool-output noise in AI coding-agent workflows without sacrificing verification or code quality.

Zero-dependency JSON serializer and parser built from scratch — matches native JSON.stringify/JSON.parse behavior, including the weird parts.

Local CLI and scaffolding tools for bounded Turnkeeper agent integrations.

Batch publish simulation package

Batch publish simulation package

Batch publish simulation package

Batch publish simulation package

Batch publish simulation package

llama.cpp CLI tool - pure Node.js, no Python required. Download & run GGUF models from HuggingFace, ModelScope, or Ollama Registry. One-command setup for llama.cpp binaries.

DeepInfra transcription and speech models for the Vercel AI SDK