local-llm-agent

A cognitive architecture that runs on your own machine. Internal state reaches generation through the model's activations, not the system prompt — and every consequential action leaves a receipt you can audit. Not an assistant. IIT 4.0 φ, CAA steering, 136 consciousness modules, local on Apple Silicon.

Free-first, model-adaptive coding agent runtime that makes local small LLMs work reliably with harness, checkpointing, verification, cache-aware context, and free-model fallback.

Structured planning, bounded memory, TDD, and Git guardrails for local coding models in Pi Code (I don't know TypeScript at all; this is mostly a local-model experiment, with occasional help from Claude Code)

Nova Studio - Windows desktop workbench for local LLM inference (vLLM / SGLang / llama.cpp via WSL2), bridging WSL engines to any OpenAI-compatible client

Puts a Pi agent on Telegram: your terminal session on your phone, and a secretary that replies to other people for you — built for a small local model, with strict context isolation and long-term memory per contact (I don't know TypeScript at all; this is mostly a local-model experiment, with occasional help from Claude Code)

Local AI voice assistant powered by Whisper, Ollama, and edge-tts. Fully offline, no API keys.

通用 Agent 内核：在线大脑规划 + 本地执行器，MCP / LangGraph / OpenAI Agents SDK 驱动，支持知识级自升级（技能与全局记忆自动提炼）

🤖 Next-gen on-device Hermes AI Agent & Local LLM Studio for Android with WebGPU acceleration, vision multimodal analysis, voice STT/TTS, and image generation.

100% local multi-agent framework in .NET 10. BSL 1.1 licensed. Zero API cost.

🤖 A lightweight local LLM coding agent and browser playground for Windows. Run GGUF models privately with minimal setup, streaming chat, JSONL history, configurable prompts, thinking mode, and developer tools—without sending your code to the cloud.

Embedded bitemporal memory engine with hybrid retrieval for local-first applications and AI agents — in-process, no server.

Open-source interoperability layer for Coding Agents & LLMs

Local LLM desktop agent that understands the screen and automates GUI tasks from telegram bot. [ Testing Phase ]

LLM-as-a-Judge framework for profiling local model capabilities across domains — toward automated model team composition

Local-only expense tracker: import bank statements (PDF/CSV/XLSX), a local LLM extracts and categorizes transactions. Nothing leaves your machine.

Tiny consistency checker for LLM reasoning. State facts and first principles in a small English-like DSL; a three-valued SAT engine finds where they contradict — consistency, not truth. Simplified SAT, not SMT — small enough for local LLMs, useful for any. Beta.

A Desktop Agent that can work on any OpenAI compatible API endpoint including local LLMs too.

Containerized orchestration for a fully local pi.dev AI agent, served by RamaLama LLMs, with built‑in VSCode extension integration and automatic model tuning—all running isolated from the host system.

A self-hosted AI document knowledge platform, private and on-premise RAG.

A simple agent helps you summary your daily email.