awesome-local-ai
152 open-source tools to run LLMs 100% locally – no cloud, no API keys, no censorship
https://github.com/ethicals7s/awesome-local-ai
Last synced: 3 days ago
JSON representation
Agent Frameworks (fully local) (25)
- AutoGen - agent system
- LangGraph - actor applications
- BabyAGI - driven autonomous agent
- Auto-GPT
- GPT Engineer
- MetaGPT - agent software company simulation
- SuperAGI
- Devon - source AI software engineer
- Open Interpreter
- Langflow
- Bisheng - code agent builder
- Taskweaver - first agent framework
- XAgent
- ChatDev
- GodMode
- Camel-AI - playing
- AgentGPT - based autonomous agents (local mode)
- Aider - aware pair programmer
Contribute
Desktop & Web UIs (22)
- Open WebUI
- LobeChat - model chat UI with local backends
- LoLLMS WebUI - in-one local LLM interface
- SillyTavern
- ComfyUI - based Stable Diffusion workflow
- InvokeAI
- Fooocus - quality image generation
- Draw Things
- Msty
- LlamaGPT - hosted chat on Umbrel
- Chatbot UI - hosted ChatGPT-like interface
- HuggingChat - hosted version of HF chat
- Taskyon - based local-first chat UI
- QA-Pilot
- Shell-Pilot - powered shell scripting
- LobeChat - model chat UI with local backends
- InvokeAI
Fine-tuning & Quantization (18)
- Unsloth - tuning on consumer GPUs
- LLaMA-Factory - tuning
- AutoGPTQ
- PEFT - efficient fine-tuning methods
- TRL
- OpenRLHF
- DeepSpeed
- Colossal-AI
- Megatron-LM
- BMTrain - efficient training
- FSDP
- LoRAX - LoRA serving
- BitsAndBytes - bit optimizers and quantization
- GPTQ-for-LLaMa - bit LLaMA quantization
- ExLlama
- ExLlamaV2
- Lit-GPT - tuning with PyTorch Lightning
- Axolotl - driven LoRA/QLoRA fine-tuning
Inference Engines & Backends (22)
- vLLM - throughput serving with PagedAttention
- TensorRT-LLM - optimized low-latency inference
- SGLang
- MLX - native framework
- MLC LLM
- ONNX Runtime - platform ML accelerator
- OpenVINO - optimized inference
- GGML
- CTranslate2
- FasterTransformer
- TurboTransformers
- LightLLM
- FlexFlow
- Triton Inference Server - framework serving
- OpenPPL
- llama.rs
One-Click Runners & Installers (15)
- text-generation-webui - rich web UI with LoRAs and extensions
- PrivateGPT
- KoboldCpp - file GGUF runner with KoboldAI API
- Tabby - hosted GitHub Copilot alternative
- Cortex - engine runner
- LMDeploy
- Llama.cpp - performance C++ inference engine (GGUF)
RAG & Vector Databases (14)
Voice & Multimodal (local) (16)
- Coqui TTS - to-speech synthesis
- OpenVoice
- Silero Models - trained TTS/STT models
- Moondream2 - language model
- Bark - to-audio with voice cloning
- Audiocraft
- RVC WebUI
- Tortoise TTS - quality multi-voice TTS
- VALL-E X - shot TTS from short audio
- Piper TTS
- OpenTTS - speaker TTS
- Kosmos-2 - text model
- ImageBind
- CLIP - image pretraining
- Whisper.cpp - to-text