#llm

Production-grade Harness Engineering Framework for multi-developer AI coding agents — guardrails, state persistence, direction alignment, and self-correction across 6 IDEs. Zero dependencies.

Your agent's most expensive failure is not forgetting. It is confidently remembering the old answer. inspeximus retires a corrected fact by key, takes the correction back on command, and leaves a receipt you can verify. Long-term memory for AI agents: one zero-dependency Python file, MCP server for Claude Code, Cursor, Windsurf, Codex and Cline.

The official Python library for the Epsilab API

Fused CUDA kernels for LLM inference - RMSNorm / RoPE / SwiGLU

Scriptable, multi-provider AI agent for the terminal — pipe into it, script it, run it on any model, cloud or local

Open-source Go agent runtime for autonomous AI systems and multi-agent automation. MCP, RAG, 15+ LLM providers and 60+ built-in tools in a single binary.

Kimi Agent SDK provides a programmatic interface to interact with the Kimi CLI

AI Badger - Local-first tool that extracts focused repo context for any AI chat (Claude, ChatGPT, Grok, etc.) without wasting tokens on irrelevant files.

Scaffold, run & deploy production LangChain agents — frontend and agent in one command. Built-in proxy, 3 chat UIs, zero CORS.

Python framework for efficient and precise interpretability research on multi-agent conversations.

The definitive TypeScript library for SpaceMolt — WebSocket-v2-first, multi-account, with live local state caches. Internals regenerate from the server's OpenAPI spec.

Terminal coding agent for local LLMs (llama.cpp, Ollama, vLLM) and any OpenAI-compatible API. OS-sandboxed autonomy, MCP, sessions. Go.

Intelligent, coordination-free loadbalancing client for Azure OpenAI

abide — type-safe isomorphic framework on Bun + web standards

No repository description published.

Enterprise engineering workflow for Claude Code — not just prompts. AI agents that plan, build, audit, and ship with security gates, privacy checks, and approval-controlled commits. Safely adopt it into new or existing repositories.

Capability Host Protocol — make agent and tool execution visible, replayable, and ready for governance.

AI agent Skill, MCP server, and CLI for transacting in the Nano (XNO) cryptocurrency. Manage wallets securely via OWS, sign messages, and build seamless zero-fee integrations.

Modular Python library for unified LLM provider access with cost reporting

Route every LLM request to the cheapest model that provably passes your quality gate — with a signed, tamper-evident receipt for every decision. Proof over prediction.

Give your AI assistant secure, read-only access to your bank accounts. MCP server supporting Plaid, Teller, Enable Banking, and Tink.

Carves your agent's tool outputs down to what matters. Never cuts what doesn't come back.

Your AI bill in one view — npx aibill. Local-first CLI that reads the Claude Code/Codex logs on your machine: real usage by project, plan value multiple, dead context, paste-ready fixes. Weekly artifacts coming: limits, wrapped, context health, secrets. Nothing leaves your laptop.

Dual-provider LLM gateway (OpenAI + Anthropic) with an agentic tool-use loop over milpa/tool-runtime. Apache-2.0.

JetStream is a throughput and memory optimized engine for LLM inference on XLA devices, starting with TPUs (and GPUs in future -- PRs welcome).

Persistent memory for AI coding agents — your agent stops forgetting between sessions. One Go binary, one SQLite file.

AI Agent SDK designed for Go developers

AI for the Command Line

Provider-agnostic, modality-aware language model router with fallback for the Vercel AI SDK.

Stop degrading your model's reasoning. A minimal, zero-config AI coding agent. Enforced ephemeral subagents keep context pure. From tiny local models up to Sol, Fable and Kimi K3.