DeepSeek

Configure DB-GPT to use DeepSeek's language models for chat and reasoning.

Prerequisites

• A DeepSeek API key
• DB-GPT installed with proxy_openai extra

Install dependencies

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

Embedding model

DeepSeek does not provide embedding models. The default config uses a HuggingFace embedding model (BAAI/bge-large-zh-v1.5). If using this, also add:

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

Configuration

Edit configs/dbgpt-proxy-deepseek.toml:

[models]
[[models.llms]]
name = "deepseek-v4-pro"
provider = "proxy/deepseek"
api_key = "your-deepseek-api-key"
# Disable V4-Pro thinking mode for strict ReAct output parsing.
thinking_enabled = false

[[models.embeddings]]
name = "BAAI/bge-large-zh-v1.5"
provider = "hf"
# Uncomment to use a local model path:
# path = "models/bge-large-zh-v1.5"

Available models

Model	Config name	Notes
DeepSeek-V4-Pro	deepseek-v4-pro	1M context, advanced reasoning, coding, and agent tasks
DeepSeek-R1	deepseek-reasoner	Strong reasoning, chain-of-thought
DeepSeek-V3	deepseek-chat	General purpose chat

For ReAct agents, keep thinking_enabled = false with deepseek-v4-pro. DeepSeek V4-Pro enables thinking mode by default, which can add reasoning blocks before the strict Thought/Action/Action Input response format.

Start the server

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

Troubleshooting

Issue	Solution
AuthenticationError	Verify your DeepSeek API key at platform.deepseek.com
Embedding download slow	Pre-download the model or use a mirror (UV_INDEX_URL)
Out of memory for embedding	Use --extra "cpu" to run embeddings on CPU

What's next

• Getting Started — Full setup walkthrough
• Model Providers — Try other providers