Skip to content

Docker Compose

The recommended way to run Heta. Brings up the full stack — API, web UI, PostgreSQL, Milvus, Neo4j, and MinIO — with a single command.

Prerequisites

  • Docker ≥ 24.0
  • Docker Compose ≥ 2.20
  • At least one LLM provider account

1. Clone and copy config

git clone https://github.com/HetaTeam/Heta.git
cd Heta
cp config.example.yaml config.yaml

git clone https://github.com/HetaTeam/Heta.git
cd Heta
cp config.example.zh.yaml config.yaml

2. Fill in API keys

The providers block in config.yaml defines service connection details. Modules reference them via YAML anchors (<<: *provider_name). The example configs use Alibaba Cloud DashScope and SiliconFlow as defaults, but any OpenAI-compatible API or self-hosted model works — just update api_key, base_url, and the model name for the relevant provider.

config.example.yaml uses four providers. Fill in all API keys — no other changes needed:

providers:
  dashscope:
    api_key: "YOUR_DASHSCOPE_API_KEY"
  siliconflow:
    api_key: "YOUR_SILICONFLOW_API_KEY"
  openai:
    api_key: "YOUR_OPENAI_API_KEY"
  gemini:
    api_key: "YOUR_GEMINI_API_KEY"

Default assignment: DashScope → HetaDB LLM/VLM and MemoryVG LLM; SiliconFlow → HetaDB embedding and HetaGen VLM/embedding; OpenAI → MemoryKB and MemoryVG embedder; Gemini → HetaGen LLM.

config.example.zh.yaml already points to DashScope and SiliconFlow with Chinese-region models. Just fill in your API keys:

providers:
  dashscope:
    api_key: "YOUR_DASHSCOPE_API_KEY"   # https://dashscope.aliyun.com
  siliconflow:
    api_key: "YOUR_SILICONFLOW_API_KEY" # https://siliconflow.cn

3. Start

docker-compose up -d

First run pulls images and builds the stack (~10–20 min).

4. Verify

docker-compose ps           # all services should show: healthy
curl localhost:8000/health

Service URLs

URL Description
http://localhost Heta web UI
http://localhost:8000/docs REST API (Swagger)
http://localhost:7474 Neo4j browser
http://localhost:9001 MinIO console

Stop

docker-compose down         # stop, keep data
docker-compose down -v      # stop and delete all volumes