DeepSeek Search

Get API Key: https://platform.deepseek.com

DeepSeek's Search API (POST https://api.deepseek.com/v1/tools/search) uses the same DEEPSEEK_API_KEY as DeepSeek chat models. See the DeepSeek Search docs for pricing and limits.

LiteLLM Python SDK

DeepSeek Search

import os
from litellm import search

os.environ["DEEPSEEK_API_KEY"] = "sk-..."

response = search(
    query="latest AI developments",
    search_provider="deepseek",
    max_results=5
)

LiteLLM AI Gateway

1. Setup config.yaml

config.yaml

model_list:
  - model_name: gpt-5.6-terra
    litellm_params:
      model: gpt-5.6-terra
      api_key: os.environ/OPENAI_API_KEY

search_tools:
  - search_tool_name: deepseek-search
    litellm_params:
      search_provider: deepseek
      api_key: os.environ/DEEPSEEK_API_KEY

2. Start the proxy

litellm --config /path/to/config.yaml

# RUNNING on http://0.0.0.0:4000

3. Test the search endpoint

Test Request

curl http://0.0.0.0:4000/v1/search/deepseek-search \
  -H "Authorization: Bearer $LITELLM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "latest AI developments",
    "max_results": 5
  }'

Parameter mapping

max_results maps to DeepSeek's count and is clamped to its 1 to 50 range. search_domain_filter is added to the query as site: operators (site:github.com OR site:huggingface.co). country and max_tokens_per_page have no DeepSeek equivalent and are dropped.

Each result carries the standard title, url, snippet and date, plus DeepSeek's summary, content, rerank_score, site_name, site_icon and authority fields as returned.

Provider-specific Parameters

DeepSeek Search with Provider-specific Parameters

import os
from litellm import search

os.environ["DEEPSEEK_API_KEY"] = "sk-..."

response = search(
    query="DeepSeek V4 release",
    search_provider="deepseek",
    max_results=5,
    # DeepSeek-specific parameters
    need_content=True,                       # Fill result.content with the page text
    ori_question="When was DeepSeek V4 released?",  # Original user question, used for reranking
    localization="en-US"                     # Preferred result language and region
)