Source Code Deployment

Deploy DB-GPT directly from source code. This is the most flexible option for development, debugging, and custom integrations.

Hardware requirements

Mode	CPU × RAM	GPU	Notes
API proxy	4C × 8 GB	None	Proxy mode does not use local GPU
Local model	8C × 32 GB	≥ 24 GB VRAM	NVIDIA GPU with CUDA support

Step 1 — Clone the repository

git clone https://github.com/eosphoros-ai/DB-GPT.git
cd DB-GPT

Step 2 — Install uv

macOS / Linux

curl -LsSf https://astral.sh/uv/install.sh | sh

PyPI (pipx)

python -m pip install --upgrade pip
python -m pip install --upgrade pipx
python -m pipx ensurepath
pipx install uv --global

Verify:

uv --version

Step 3 — Install dependencies

OpenAI (proxy)

uv sync --all-packages \
  --extra "base" \
  --extra "proxy_openai" \
  --extra "rag" \
  --extra "storage_chromadb" \
  --extra "dbgpts"

DeepSeek (proxy)

uv sync --all-packages \
  --extra "base" \
  --extra "proxy_openai" \
  --extra "rag" \
  --extra "storage_chromadb" \
  --extra "dbgpts"

info

DeepSeek uses the OpenAI-compatible proxy, so the extras are the same as OpenAI.

Ollama (local)

uv sync --all-packages \
  --extra "base" \
  --extra "proxy_ollama" \
  --extra "rag" \
  --extra "storage_chromadb" \
  --extra "dbgpts"

Local GPU (HuggingFace)

uv sync --all-packages \
  --extra "base" \
  --extra "cuda121" \
  --extra "hf" \
  --extra "rag" \
  --extra "storage_chromadb" \
  --extra "quant_bnb" \
  --extra "dbgpts"

Use the interactive install helper

DB-GPT provides an interactive helper to generate the right uv sync command:

uv run install_help.py install-cmd --interactive

Or list all available extras:

uv run install_help.py list

Step 4 — Configure your model

Edit the TOML config file for your chosen provider. See Model Providers for details.

OpenAI

Edit configs/dbgpt-proxy-openai.toml:

[models]
[[models.llms]]
name = "chatgpt_proxyllm"
provider = "proxy/openai"
api_key = "your-openai-api-key"    # <-- replace

[[models.embeddings]]
name = "text-embedding-3-small"
provider = "proxy/openai"
api_key = "your-openai-api-key"    # <-- replace

DeepSeek

Edit configs/dbgpt-proxy-deepseek.toml:

[models]
[[models.llms]]
name = "deepseek-reasoner"
provider = "proxy/deepseek"
api_key = "your-deepseek-api-key"  # <-- replace

[[models.embeddings]]
name = "BAAI/bge-large-zh-v1.5"
provider = "hf"

info

If using a HuggingFace embedding, also add --extra "hf" and --extra "cpu" to the install command.

Ollama

Make sure Ollama is running, then edit configs/dbgpt-proxy-ollama.toml:

[models]
[[models.llms]]
name = "qwen2.5:latest"
provider = "proxy/ollama"
api_base = "http://localhost:11434"

[[models.embeddings]]
name = "nomic-embed-text:latest"
provider = "proxy/ollama"
api_base = "http://localhost:11434"

Environment variables

Use "${env:OPENAI_API_KEY}" syntax in TOML to read from environment variables instead of hardcoding keys.

Step 5 — Start the server

OpenAI

uv run dbgpt start webserver --config configs/dbgpt-proxy-openai.toml

DeepSeek

uv run dbgpt start webserver --config configs/dbgpt-proxy-deepseek.toml

Ollama

uv run dbgpt start webserver --config configs/dbgpt-proxy-ollama.toml

Step 6 — Open the Web UI

Open your browser and visit http://localhost:5670.

Verify it works

If the Web UI loads and you can start a chat conversation, your DB-GPT is running.

Common first-run issues

• uv sync fails
  • Re-check Python and uv: Prerequisites
  • If you are in China, use a mirror via UV_INDEX_URL
• Provider auth fails
  • Verify the selected TOML file under configs/
  • Check the matching provider guide: Model Providers
• Server starts but UI is blank
  • Confirm the terminal shows the webserver started cleanly
  • Check whether another process is already using port 5670

Database configuration

SQLite (default)

SQLite is the default — tables are created automatically. No extra setup needed.

[service.web.database]
type = "sqlite"
path = "pilot/meta_data/dbgpt.db"

MySQL

1. Create the database:

mysql -h127.0.0.1 -uroot -p{your_password} < ./assets/schema/dbgpt.sql

2. Update your TOML config:

[service.web.database]
type = "mysql"
host = "127.0.0.1"
port = 3306
user = "root"
database = "dbgpt"
password = "your-password"

Load test data (optional)

# Linux / macOS
bash ./scripts/examples/load_examples.sh

# Windows
.\scripts\examples\load_examples.bat

Run web front-end separately (optional)

For front-end development or custom UI work:

cd web && npm install
cp .env.template .env
# Edit .env — set API_BASE_URL=http://localhost:5670
npm run dev

Open http://localhost:3000.

Next steps

Topic	Link
Configure more model providers	Model Providers
Deploy with Docker	Docker
Deploy as a cluster	Cluster
Explore the Web UI	Web UI Guide