Micro Models Agent (MMA) — LLM agent harness for small models (Qwen3.5-9B, 32K-64K context)

Give brains to your small models. MCP server enforcing step-by-step Chain-of-Thought.

Agentic CLI for local LLMs (LM Studio, Ollama, llama.cpp). Real agent loop, web search, skill memory, auto-compaction, MCP server. Optimized for small models (4-14B).

Tiny TypeScript-native tensor library with autograd, compiling to WebGPU. Train small models in the browser without hand-writing kernels, or run a pretrained one frozen for transfer learning.

MSM — Multi Small Models: An open standard for coordinated small model pipelines

Lint and rewrite AI harness markdown so weak small models (Minimax 2.5, Nemotron 3, Mistral 7B) can execute it reliably.

Lint and rewrite AI harness markdown so weak small models (Minimax 2.5, Nemotron 3, Mistral 7B) can execute it reliably.

Tree-shakeable static models.dev catalog split by provider for TokenLens.

NodeJS implementation of @Qdrant/fastembed

Local embedding model integration for Mastra, powered by ONNX Runtime.

Hardware inspector for Avasis Agent Builder. Detects CPU, RAM, GPU and recommends small models that run on your device.

TUF metadata models

MistralAI integration for LangChain.js

Use Custom Media Queries in CSS

Contains JavaScript & TypeScript object models for Microsoft Power BI JavaScript SDK. For each model there is a TypeScript interface, and a validation function to ensure and object is valid.

Manage state with style in every framework

Mintlify models

A Javascript library for running ONNX models on browsers

A small, framework-agnostic JavaScript/TypeScript client for the Apache Iceberg REST Catalog

An open-source machine learning framework.