The DeepSeek layer
for your stack

Multi-turn contextOpenAI

Send full conversation history in messages — system + prior user/assistant turns. The model sees the whole thread.

{
  "model": "deepseek-chat",
  "messages": [
    { "role": "system", "content": "You help refactor Python code. Be concise." },
    { "role": "user", "content": "Here is my auth module. How can I simplify it?" },
    { "role": "assistant", "content": "You can extract token validation into a dependency..." },
    { "role": "user", "content": "Show the dependency as code." }
  ],
  "max_tokens": 4096
}

system   — instructions / persona (optional)
user     — human or client message
assistant — previous model replies (include for context)
tool     — tool result after tool_calls (agents / function calling)

{
  "model": "deepseek-chat",
  "messages": [
    { "role": "user", "content": "What is the weather in Berlin?" },
    {
      "role": "assistant",
      "content": null,
      "tool_calls": [{
        "id": "call_1",
        "type": "function",
        "function": { "name": "get_weather", "arguments": "{\"city\":\"Berlin\"}" }
      }]
    },
    { "role": "tool", "tool_call_id": "call_1", "content": "{\"temp_c\": 12, \"condition\": \"cloudy\"}" }
  ]
}

Web search modelslive

Use *-search model IDs for web-enabled replies. Vision (deepseek-vision) is not available right now — returns 400.

{
  "model": "deepseek-chat-search",
  "messages": [
    { "role": "user", "content": "What changed in Python 3.13? Cite sources briefly." }
  ],
  "max_tokens": 2048
}

deepseek-chat-search
deepseek-default-search
deepseek-reasoner-search
deepseek-r1-search

Tools / function callingverified live

Works. Send OpenAI tools JSON — response returns tool_calls + finish_reason: "tool_calls". Tested on production with deepseek-chat, deepseek-reasoner, deepseek-v4-pro, deepseek-default, deepseek-expert. Cline / Continue agents use this path.

curl https://deepseek.llm-api.fun/v1/chat/completions \
  -H "Authorization: Bearer dllm-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "messages": [
      { "role": "user", "content": "Call read_file for path config.json only. Do not explain." }
    ],
    "tools": [{
      "type": "function",
      "function": {
        "name": "read_file",
        "description": "Read a file from disk",
        "parameters": {
          "type": "object",
          "properties": {
            "path": { "type": "string", "description": "File path" }
          },
          "required": ["path"]
        }
      }
    }],
    "tool_choice": "auto",
    "max_tokens": 512
  }'

{
  "model": "deepseek-chat",
  "choices": [{
    "message": {
      "role": "assistant",
      "content": null,
      "tool_calls": [{
        "id": "call_1781005249449_n5addi",
        "type": "function",
        "function": {
          "name": "read_file",
          "arguments": "{\"path\":\"config.json\"}"
        }
      }]
    },
    "finish_reason": "tool_calls"
  }]
}

{
  "model": "deepseek-reasoner",
  "messages": [{ "role": "user", "content": "Call get_weather for Berlin." }],
  "tools": [{
    "type": "function",
    "function": {
      "name": "get_weather",
      "description": "Get weather for a city",
      "parameters": {
        "type": "object",
        "properties": { "city": { "type": "string" } },
        "required": ["city"]
      }
    }
  }],
  "tool_choice": "auto"
}

Anthropic MessagesPOST

Same key — Anthropic-style body. Use when your client sends x-api-key instead of Bearer.

curl https://deepseek.llm-api.fun/v1/messages \
  -H "x-api-key: dllm-YOUR_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "max_tokens": 4096,
    "system": "You are a coding assistant.",
    "messages": [
      { "role": "user", "content": "Refactor this function to use async/await." }
    ]
  }'

{
  "model": "deepseek-chat",
  "max_tokens": 4096,
  "system": "You are a helpful agent.",
  "tools": [{
    "name": "get_weather",
    "description": "Get weather for a city",
    "input_schema": {
      "type": "object",
      "properties": { "city": { "type": "string" } },
      "required": ["city"]
    }
  }],
  "messages": [
    { "role": "user", "content": "Weather in Berlin?" },
    {
      "role": "assistant",
      "content": [{
        "type": "tool_use",
        "id": "tu_1",
        "name": "get_weather",
        "input": { "city": "Berlin" }
      }]
    },
    {
      "role": "user",
      "content": [{
        "type": "tool_result",
        "tool_use_id": "tu_1",
        "content": "{\"temp_c\": 12}"
      }]
    }
  ]
}

Streaming (SSE)stream: true

Set "stream": true on OpenAI or Anthropic requests. Response is text/event-stream with data: {...} chunks.

curl https://deepseek.llm-api.fun/v1/chat/completions \
  -H "Authorization: Bearer dllm-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-reasoner",
    "stream": true,
    "messages": [{ "role": "user", "content": "Explain recursion briefly" }]
  }'

from openai import OpenAI

client = OpenAI(
    api_key="dllm-YOUR_KEY",
    base_url="https://deepseek.llm-api.fun/v1",
)
stream = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "Hello"}],
    stream=True,
)
for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

Discover models

curl https://deepseek.llm-api.fun/v1/models \
  -H "Authorization: Bearer dllm-YOUR_KEY"

curl https://deepseek.llm-api.fun/v1/model-capabilities \
  -H "Authorization: Bearer dllm-YOUR_KEY"

Use any ID from the response as model in chat requests. Live list also on the models section above.