Serge is a chat interface crafted with llama.cpp for running Alpaca models. No API keys, entirely self-hosted!

• 🌐 SvelteKit frontend
• 💾 Redis for storing chat history & parameters
• ⚙️ FastAPI + LangChain for the API, wrapping calls to llama.cpp using the python bindings

🎥 Demo:

🐳 Docker:

docker run -d \
    --name serge \
    -v weights:/usr/src/app/weights \
    -v datadb:/data/db/ \
    -p 8008:8008 \
    ghcr.io/serge-chat/serge:latest

🐙 Docker Compose:

services:
  serge:
    image: ghcr.io/serge-chat/serge:latest
    container_name: serge
    restart: unless-stopped
    ports:
      - 8008:8008
    volumes:
      - weights:/usr/src/app/weights
      - datadb:/data/db/

volumes:
  weights:
  datadb:

Then, just visit http://localhost:8008/, You can find the API documentation at http://localhost:8008/api/docs

Ensure you have Docker Desktop installed, WSL2 configured, and enough free RAM to run models.

Instructions for setting up Serge on Kubernetes can be found in the wiki.

We currently support the following models:

• Airoboros 🎈
  • Airoboros-7B
  • Airoboros-13B
  • Airoboros-30B
• Alpaca 🦙
  • Alpaca-LoRA-65B
  • GPT4-Alpaca-LoRA-30B
• Chronos 🌑
  • Chronos-13B
  • Chronos-33B
• GPT4All 🌍
  • GPT4All-13B
• Guanaco 🦙
  • Guanaco-7B
  • Guanaco-13B
  • Guanaco-33B
  • Guanaco-65B
• Koala 🐨
  • Koala-7B
  • Koala-13B
• Llama 🦙
  • FinLlama-33B
  • Llama-Supercot-30B
• Lazarus 💀
  • Lazarus-30B
• Nous 🧠
  • Nous-Hermes-13B
• OpenAssistant 🎙️
  • OpenAssistant-30B
• Samantha 👩
  • Samantha-7B
  • Samantha-13B
  • Samantha-33B
• Tulu 🎚
  • Tulu-7B
  • Tulu-13B
  • Tulu-30B
• Vicuna 🦙
  • Stable-Vicuna-13B
  • Vicuna-CoT-7B
  • Vicuna-CoT-13B
  • Vicuna-v1.1-7B
  • Vicuna-v1.1-13B
  • VicUnlocked-30B
  • VicUnlocked-65B
• Wizard 🧙
  • Wizard-Mega-13B
  • Wizard-Vicuna-Uncensored-7B
  • Wizard-Vicuna-Uncensored-13B
  • Wizard-Vicuna-Uncensored-30B
  • WizardLM-30B
  • WizardLM-Uncensored-7B
  • WizardLM-Uncensored-13B
  • WizardLM-Uncensored-30B

Additional weights can be added to the serge_weights volume using docker cp:

docker cp ./my_weight.bin serge:/usr/src/app/weights/

LLaMA will crash if you don't have enough available memory for the model:

Need help? Join our Discord

If you discover a bug or have a feature idea, feel free to open an issue or PR.

To run Serge in development mode:

git clone https://github.com/serge-chat/serge.git
DOCKER_BUILDKIT=1 docker compose -f docker-compose.dev.yml up -d --build