Skip to content

Use Claude Code, Codex, VSCode, Pi, and OpenCode (and 6 other harnesses) for free (1.3B+ free tokens) from your terminal, app, IDE, or phone, and now from the browser with native browser sessions (multi-harness + multi-model) like OpenClaw (voice supported + ToS friendly)

Python

feat: deepseek api support (#118)

Pavel Yurchenkoe719e4aed2b6
diff --git a/.env.example b/.env.example
index cd0169f7..c97a7bd3 100644
--- a/.env.example
+++ b/.env.example
@@ -6,6 +6,10 @@ NVIDIA_NIM_API_KEY=""
OPENROUTER_API_KEY=""
+# DeepSeek Config
+DEEPSEEK_API_KEY=""
+
+
# LM Studio Config (local provider, no API key required)
LM_STUDIO_BASE_URL="http://localhost:1234/v1"
@@ -16,7 +20,7 @@ LLAMACPP_BASE_URL="http://localhost:8080/v1"
# All Claude model requests are mapped to these models, plain model is fallback
# Format: provider_type/model/name
-# Valid providers: "nvidia_nim" | "open_router" | "lmstudio" | "llamacpp"
+# Valid providers: "nvidia_nim" | "open_router" | "deepseek" | "lmstudio" | "llamacpp"
MODEL_OPUS="nvidia_nim/z-ai/glm4.7"
MODEL_SONNET="open_router/arcee-ai/trinity-large-preview:free"
MODEL_HAIKU="open_router/stepfun/step-3.5-flash:free"
diff --git a/README.md b/README.md
index db09a91d..76ea6cab 100644
--- a/README.md
+++ b/README.md
@@ -12,7 +12,7 @@
[![Code style: Ruff](https://img.shields.io/badge/code%20formatting-ruff-f5a623.svg?style=for-the-badge)](https://github.com/astral-sh/ruff)
[![Logging: Loguru](https://img.shields.io/badge/logging-loguru-4ecdc4.svg?style=for-the-badge)](https://github.com/Delgan/loguru)
-A lightweight proxy that routes Claude Code's Anthropic API calls to **NVIDIA NIM** (40 req/min free), **OpenRouter** (hundreds of models), **LM Studio** (fully local), or **llama.cpp** (local with Anthropic endpoints).
+A lightweight proxy that routes Claude Code's Anthropic API calls to **NVIDIA NIM** (40 req/min free), **OpenRouter** (hundreds of models), **DeepSeek** (direct API), **LM Studio** (fully local), or **llama.cpp** (local with Anthropic endpoints).
[Quick Start](#quick-start) · [Providers](#providers) · [Discord Bot](#discord-bot) · [Configuration](#configuration) · [Development](#development) · [Contributing](#contributing)
@@ -31,7 +31,7 @@ A lightweight proxy that routes Claude Code's Anthropic API calls to **NVIDIA NI
| -------------------------- | ----------------------------------------------------------------------------------------------- |
| **Zero Cost** | 40 req/min free on NVIDIA NIM. Free models on OpenRouter. Fully local with LM Studio |
| **Drop-in Replacement** | Set 2 env vars. No modifications to Claude Code CLI or VSCode extension needed |
-| **4 Providers** | NVIDIA NIM, OpenRouter (hundreds of models), LM Studio (local), llama.cpp (`llama-server`) |
+| **5 Providers** | NVIDIA NIM, OpenRouter, DeepSeek, LM Studio (local), llama.cpp (`llama-server`) |
| **Per-Model Mapping** | Route Opus / Sonnet / Haiku to different models and providers. Mix providers freely |
| **Thinking Token Support** | Parses `<think>` tags and `reasoning_content` into native Claude thinking blocks |
| **Heuristic Tool Parser** | Models outputting tool calls as text are auto-parsed into structured tool use |
@@ -48,6 +48,7 @@ A lightweight proxy that routes Claude Code's Anthropic API calls to **NVIDIA NI
1. Get an API key (or use LM Studio / llama.cpp locally):
- **NVIDIA NIM**: [build.nvidia.com/settings/api-keys](https://build.nvidia.com/settings/api-keys)
- **OpenRouter**: [openrouter.ai/keys](https://openrouter.ai/keys)
+ - **DeepSeek**: [platform.deepseek.com/api_keys](https://platform.deepseek.com/api_keys)
- **LM Studio**: No API key needed. Run locally with [LM Studio](https://lmstudio.ai)
- **llama.cpp**: No API key needed. Run `llama-server` locally.
2. Install [Claude Code](https://github.com/anthropics/claude-code)
@@ -100,6 +101,20 @@ MODEL="open_router/stepfun/step-3.5-flash:free" # fallback
</details>
+<details>
+<summary><b>DeepSeek</b> (direct API)</summary>
+
+```dotenv
+DEEPSEEK_API_KEY="your-deepseek-key-here"
+
+MODEL_OPUS="deepseek/deepseek-reasoner"
+MODEL_SONNET="deepseek/deepseek-chat"
+MODEL_HAIKU="deepseek/deepseek-chat"
+MODEL="deepseek/deepseek-chat" # fallback
+```
+
+</details>
+
<details>
<summary><b>LM Studio</b> (fully local, no API key)</summary>
@@ -293,6 +308,7 @@ The proxy also exposes Claude-compatible probe routes: `GET /v1/models`, `POST /
| -------------- | ------------ | ---------- | ------------------------------------ |
| **NVIDIA NIM** | Free | 40 req/min | Daily driver, generous free tier |
| **OpenRouter** | Free / Paid | Varies | Model variety, fallback options |
+| **DeepSeek** | Usage-based | Varies | Direct access to DeepSeek chat/reasoner |
| **LM Studio** | Free (local) | Unlimited | Privacy, offline use, no rate limits |
| **llama.cpp** | Free (local) | Unlimited | Lightweight local inference engine |
@@ -302,6 +318,7 @@ Models use a prefix format: `provider_prefix/model/name`. An invalid prefix caus
| ---------- | ----------------- | -------------------- | ----------------------------- |
| NVIDIA NIM | `nvidia_nim/...` | `NVIDIA_NIM_API_KEY` | `integrate.api.nvidia.com/v1` |
| OpenRouter | `open_router/...` | `OPENROUTER_API_KEY` | `openrouter.ai/api/v1` |
+| DeepSeek | `deepseek/...` | `DEEPSEEK_API_KEY` | `api.deepseek.com` |
| LM Studio | `lmstudio/...` | (none) | `localhost:1234/v1` |
| llama.cpp | `llamacpp/...` | (none) | `localhost:8080/v1` |
@@ -334,6 +351,18 @@ Browse: [openrouter.ai/models](https://openrouter.ai/models) · [Free models](ht
</details>
+<details>
+<summary><b>DeepSeek models</b></summary>
+
+DeepSeek currently exposes the direct API models:
+
+- `deepseek/deepseek-chat`
+- `deepseek/deepseek-reasoner`
+
+Browse: [api-docs.deepseek.com](https://api-docs.deepseek.com)
+
+</details>
+
<details>
<summary><b>LM Studio models</b></summary>
@@ -455,6 +484,7 @@ Configure via `WHISPER_DEVICE` (`cpu` | `cuda` | `nvidia_nim`) and `WHISPER_MODE
| `NVIDIA_NIM_API_KEY` | NVIDIA API key | required for NIM |
| `ENABLE_THINKING` | Global switch for provider reasoning requests and Claude thinking blocks. Set `false` to hide thinking across all providers. | `true` |
| `OPENROUTER_API_KEY` | OpenRouter API key | required for OpenRouter |
+| `DEEPSEEK_API_KEY` | DeepSeek API key | required for DeepSeek |
| `LM_STUDIO_BASE_URL` | LM Studio server URL | `http://localhost:1234/v1` |
| `LLAMACPP_BASE_URL` | llama.cpp server URL | `http://localhost:8080/v1` |
@@ -514,7 +544,7 @@ See [`.env.example`](.env.example) for all supported parameters.
free-claude-code/
├── server.py # Entry point
├── api/ # FastAPI routes, request detection, optimization handlers
-├── providers/ # BaseProvider, OpenAICompatibleProvider, NIM, OpenRouter, LM Studio, llamacpp
+├── providers/ # BaseProvider, OpenAICompatibleProvider, NIM, OpenRouter, DeepSeek, LM Studio, llamacpp
│ └── common/ # Shared utils (SSE builder, message converter, parsers, error mapping)
├── messaging/ # MessagingPlatform ABC + Discord/Telegram bots, session management
├── config/ # Settings, NIM config, logging
diff --git a/api/dependencies.py b/api/dependencies.py
index 40a962e9..1585ca45 100644
--- a/api/dependencies.py
+++ b/api/dependencies.py
@@ -7,6 +7,7 @@ from config.settings import Settings
from config.settings import get_settings as _get_settings
from providers.base import BaseProvider, ProviderConfig
from providers.common import get_user_facing_error_message
+from providers.deepseek import DEEPSEEK_BASE_URL, DeepSeekProvider
from providers.exceptions import AuthenticationError
from providers.llamacpp import LlamaCppProvider
from providers.lmstudio import LMStudioProvider
@@ -60,6 +61,24 @@ def _create_provider_for_type(provider_type: str, settings: Settings) -> BasePro
enable_thinking=settings.enable_thinking,
)
return OpenRouterProvider(config)
+ if provider_type == "deepseek":
+ if not settings.deepseek_api_key or not settings.deepseek_api_key.strip():
+ raise AuthenticationError(
+ "DEEPSEEK_API_KEY is not set. Add it to your .env file. "
+ "Get a key at https://platform.deepseek.com/api_keys"
+ )
+ config = ProviderConfig(
+ api_key=settings.deepseek_api_key,
+ base_url=DEEPSEEK_BASE_URL,
+ rate_limit=settings.provider_rate_limit,
+ rate_window=settings.provider_rate_window,
+ max_concurrency=settings.provider_max_concurrency,
+ http_read_timeout=settings.http_read_timeout,
+ http_write_timeout=settings.http_write_timeout,
+ http_connect_timeout=settings.http_connect_timeout,
+ enable_thinking=settings.enable_thinking,
+ )
+ return DeepSeekProvider(config)
if provider_type == "lmstudio":
config = ProviderConfig(
api_key="lm-studio",
@@ -87,12 +106,12 @@ def _create_provider_for_type(provider_type: str, settings: Settings) -> BasePro
)
return LlamaCppProvider(config)
logger.error(
- "Unknown provider_type: '{}'. Supported: 'nvidia_nim', 'open_router', 'lmstudio', 'llamacpp'",
+ "Unknown provider_type: '{}'. Supported: 'nvidia_nim', 'open_router', 'deepseek', 'lmstudio', 'llamacpp'",
provider_type,
)
raise ValueError(
f"Unknown provider_type: '{provider_type}'. "
- f"Supported: 'nvidia_nim', 'open_router', 'lmstudio', 'llamacpp'"
+ f"Supported: 'nvidia_nim', 'open_router', 'deepseek', 'lmstudio', 'llamacpp'"
)
diff --git a/config/env.example b/config/env.example
index 8965e367..a77fd4e1 100644
--- a/config/env.example
+++ b/config/env.example
@@ -6,13 +6,17 @@ NVIDIA_NIM_API_KEY=""
OPENROUTER_API_KEY=""
+# DeepSeek Config
+DEEPSEEK_API_KEY=""
+
+
# LM Studio Config (local provider, no API key required)
LM_STUDIO_BASE_URL="http://localhost:1234/v1"
# All Claude model requests are mapped to these models, plain model is fallback
# Format: provider_type/model/name
-# Valid providers: "nvidia_nim" | "open_router" | "lmstudio"
+# Valid providers: "nvidia_nim" | "open_router" | "deepseek" | "lmstudio" | "llamacpp"
MODEL_OPUS="nvidia_nim/z-ai/glm4.7"
MODEL_SONNET="open_router/arcee-ai/trinity-large-preview:free"
MODEL_HAIKU="open_router/stepfun/step-3.5-flash:free"
@@ -68,4 +72,4 @@ FAST_PREFIX_DETECTION=true
ENABLE_NETWORK_PROBE_MOCK=true
ENABLE_TITLE_GENERATION_SKIP=true
ENABLE_SUGGESTION_MODE_SKIP=true
-ENABLE_FILEPATH_EXTRACTION_MOCK=true
\ No newline at end of file
+ENABLE_FILEPATH_EXTRACTION_MOCK=true
diff --git a/config/settings.py b/config/settings.py
index 6e2fdc6b..f0d8d49e 100644
--- a/config/settings.py
+++ b/config/settings.py
@@ -81,6 +81,9 @@ class Settings(BaseSettings):
# ==================== OpenRouter Config ====================
open_router_api_key: str = Field(default="", validation_alias="OPENROUTER_API_KEY")
+ # ==================== DeepSeek Config ====================
+ deepseek_api_key: str = Field(default="", validation_alias="DEEPSEEK_API_KEY")
+
# ==================== Messaging Platform Selection ====================
# Valid: "telegram" | "discord"
messaging_platform: str = Field(
@@ -219,7 +222,13 @@ class Settings(BaseSettings):
def validate_model_format(cls, v: str | None) -> str | None:
if v is None:
return None
- valid_providers = ("nvidia_nim", "open_router", "lmstudio", "llamacpp")
+ valid_providers = (
+ "nvidia_nim",
+ "open_router",
+ "deepseek",
+ "lmstudio",
+ "llamacpp",
+ )
if "/" not in v:
raise ValueError(
f"Model must be prefixed with provider type. "
@@ -230,7 +239,7 @@ class Settings(BaseSettings):
if provider not in valid_providers:
raise ValueError(
f"Invalid provider: '{provider}'. "
- f"Supported: 'nvidia_nim', 'open_router', 'lmstudio', 'llamacpp'"
+ f"Supported: 'nvidia_nim', 'open_router', 'deepseek', 'lmstudio', 'llamacpp'"
)
return v
diff --git a/providers/__init__.py b/providers/__init__.py
index e6a759e5..c9206235 100644
--- a/providers/__init__.py
+++ b/providers/__init__.py
@@ -1,6 +1,7 @@
"""Providers package - implement your own provider by extending BaseProvider."""
from .base import BaseProvider, ProviderConfig
+from .deepseek import DeepSeekProvider
from .exceptions import (
APIError,
AuthenticationError,
@@ -18,6 +19,7 @@ __all__ = [
"APIError",
"AuthenticationError",
"BaseProvider",
+ "DeepSeekProvider",
"InvalidRequestError",
"LMStudioProvider",
"LlamaCppProvider",
diff --git a/providers/common/message_converter.py b/providers/common/message_converter.py
index d2b2881f..2e9e5933 100644
--- a/providers/common/message_converter.py
+++ b/providers/common/message_converter.py
@@ -27,11 +27,12 @@ class AnthropicToOpenAIConverter:
*,
include_thinking: bool = True,
include_reasoning_for_openrouter: bool = False,
+ include_reasoning_content: bool = False,
) -> list[dict[str, Any]]:
"""Convert a list of Anthropic messages to OpenAI format.
- When include_reasoning_for_openrouter is True, assistant messages with
- thinking blocks get reasoning_content added for OpenRouter multi-turn
+ When reasoning_content preservation is enabled, assistant messages with
+ thinking blocks get reasoning_content added for provider multi-turn
reasoning continuation.
"""
result = []
@@ -49,6 +50,7 @@ class AnthropicToOpenAIConverter:
content,
include_thinking=include_thinking,
include_reasoning_for_openrouter=include_reasoning_for_openrouter,
+ include_reasoning_content=include_reasoning_content,
)
)
elif role == "user":
@@ -66,11 +68,15 @@ class AnthropicToOpenAIConverter:
*,
include_thinking: bool = True,
include_reasoning_for_openrouter: bool = False,
+ include_reasoning_content: bool = False,
) -> list[dict[str, Any]]:
"""Convert assistant message blocks, preserving interleaved thinking+text order."""
content_parts: list[str] = []
thinking_parts: list[str] = []
tool_calls: list[dict[str, Any]] = []
+ emit_reasoning_content = (
+ include_reasoning_for_openrouter or include_reasoning_content
+ )
for block in content:
block_type = get_block_type(block)
@@ -82,7 +88,7 @@ class AnthropicToOpenAIConverter:
continue
thinking = get_block_attr(block, "thinking", "")
content_parts.append(f"<think>\n{thinking}\n</think>")
- if include_reasoning_for_openrouter:
+ if emit_reasoning_content:
thinking_parts.append(thinking)
elif block_type == "tool_use":
tool_input = get_block_attr(block, "input", {})
@@ -112,7 +118,7 @@ class AnthropicToOpenAIConverter:
}
if tool_calls:
msg["tool_calls"] = tool_calls
- if include_reasoning_for_openrouter and thinking_parts:
+ if emit_reasoning_content and thinking_parts:
msg["reasoning_content"] = "\n".join(thinking_parts)
return [msg]
@@ -191,6 +197,7 @@ def build_base_request_body(
default_max_tokens: int | None = None,
include_thinking: bool = True,
include_reasoning_for_openrouter: bool = False,
+ include_reasoning_content: bool = False,
) -> dict[str, Any]:
"""Build the common parts of an OpenAI-format request body.
@@ -204,6 +211,7 @@ def build_base_request_body(
request_data.messages,
include_thinking=include_thinking,
include_reasoning_for_openrouter=include_reasoning_for_openrouter,
+ include_reasoning_content=include_reasoning_content,
)
system = getattr(request_data, "system", None)
diff --git a/providers/deepseek/__init__.py b/providers/deepseek/__init__.py
new file mode 100644
index 00000000..0ecf7d05
--- /dev/null
+++ b/providers/deepseek/__init__.py
@@ -0,0 +1,5 @@
+"""DeepSeek provider exports."""
+
+from .client import DEEPSEEK_BASE_URL, DeepSeekProvider
+
+__all__ = ["DEEPSEEK_BASE_URL", "DeepSeekProvider"]
diff --git a/providers/deepseek/client.py b/providers/deepseek/client.py
new file mode 100644
index 00000000..56a7ade1
--- /dev/null
+++ b/providers/deepseek/client.py
@@ -0,0 +1,29 @@
+"""DeepSeek provider implementation."""
+
+from typing import Any
+
+from providers.base import ProviderConfig
+from providers.openai_compat import OpenAICompatibleProvider
+
+from .request import build_request_body
+
+DEEPSEEK_BASE_URL = "https://api.deepseek.com"
+
+
+class DeepSeekProvider(OpenAICompatibleProvider):
+ """DeepSeek provider using OpenAI-compatible chat completions."""
+
+ def __init__(self, config: ProviderConfig):
+ super().__init__(
+ config,
+ provider_name="DEEPSEEK",
+ base_url=config.base_url or DEEPSEEK_BASE_URL,
+ api_key=config.api_key,
+ )
+
+ def _build_request_body(self, request: Any) -> dict:
+ """Internal helper for tests and shared building."""
+ return build_request_body(
+ request,
+ thinking_enabled=self._is_thinking_enabled(request),
+ )
diff --git a/providers/deepseek/request.py b/providers/deepseek/request.py
new file mode 100644
index 00000000..e274be7d
--- /dev/null
+++ b/providers/deepseek/request.py
@@ -0,0 +1,39 @@
+"""Request builder for DeepSeek provider."""
+
+from typing import Any
+
+from loguru import logger
+
+from providers.common.message_converter import build_base_request_body
+
+
+def build_request_body(request_data: Any, *, thinking_enabled: bool) -> dict:
+ """Build OpenAI-format request body from Anthropic request for DeepSeek."""
+ logger.debug(
+ "DEEPSEEK_REQUEST: conversion start model={} msgs={}",
+ getattr(request_data, "model", "?"),
+ len(getattr(request_data, "messages", [])),
+ )
+ body = build_base_request_body(
+ request_data,
+ include_reasoning_content=True,
+ )
+
+ extra_body: dict[str, Any] = {}
+ request_extra = getattr(request_data, "extra_body", None)
+ if request_extra:
+ extra_body.update(request_extra)
+
+ if thinking_enabled and body.get("model") != "deepseek-reasoner":
+ extra_body.setdefault("thinking", {"type": "enabled"})
+
+ if extra_body:
+ body["extra_body"] = extra_body
+
+ logger.debug(
+ "DEEPSEEK_REQUEST: conversion done model={} msgs={} tools={}",
+ body.get("model"),
+ len(body.get("messages", [])),
+ len(body.get("tools", [])),
+ )
+ return body
diff --git a/tests/api/test_dependencies.py b/tests/api/test_dependencies.py
index bd30b95b..ebf657c8 100644
--- a/tests/api/test_dependencies.py
+++ b/tests/api/test_dependencies.py
@@ -10,6 +10,7 @@ from api.dependencies import (
get_settings,
)
from config.nim import NimSettings
+from providers.deepseek import DeepSeekProvider
from providers.lmstudio import LMStudioProvider
from providers.nvidia_nim import NvidiaNimProvider
from providers.open_router import OpenRouterProvider
@@ -25,11 +26,13 @@ def _make_mock_settings(**overrides):
mock.provider_rate_window = 60
mock.provider_max_concurrency = 5
mock.open_router_api_key = "test_openrouter_key"
+ mock.deepseek_api_key = "test_deepseek_key"
mock.lm_studio_base_url = "http://localhost:1234/v1"
mock.nim = NimSettings()
mock.http_read_timeout = 300.0
mock.http_write_timeout = 10.0
mock.http_connect_timeout = 2.0
+ mock.enable_thinking = True
for key, value in overrides.items():
setattr(mock, key, value)
return mock
@@ -120,6 +123,49 @@ async def test_get_provider_lmstudio():
assert provider._base_url == "http://localhost:1234/v1"
+async def test_get_provider_deepseek():
+ """Test that provider_type=deepseek returns DeepSeekProvider."""
+ with patch("api.dependencies.get_settings") as mock_settings:
+ mock_settings.return_value = _make_mock_settings(provider_type="deepseek")
+
+ provider = get_provider()
+
+ assert isinstance(provider, DeepSeekProvider)
+ assert provider._base_url == "https://api.deepseek.com"
+ assert provider._api_key == "test_deepseek_key"
+ assert provider._config.enable_thinking is True
+
+
+async def test_get_provider_deepseek_uses_fixed_base_url():
+ """DeepSeek provider always uses the fixed provider base URL."""
+ with patch("api.dependencies.get_settings") as mock_settings:
+ mock_settings.return_value = _make_mock_settings(
+ provider_type="deepseek",
+ )
+
+ provider = get_provider()
+
+ assert isinstance(provider, DeepSeekProvider)
+ assert provider._base_url == "https://api.deepseek.com"
+
+
+async def test_get_provider_deepseek_passes_enable_thinking():
+ """DeepSeek provider receives the global thinking toggle."""
+ with patch("api.dependencies.get_settings") as mock_settings:
+ mock_settings.return_value = _make_mock_settings(
+ provider_type="deepseek",
+ enable_thinking=False,
+ )
+
+ provider = get_provider()
+
+ assert isinstance(provider, DeepSeekProvider)
+ assert provider._config.enable_thinking is False
+
+
@pytest.mark.asyncio
async def test_get_provider_lmstudio_uses_lm_studio_base_url():
"""LM Studio provider uses lm_studio_base_url from settings."""
@@ -200,6 +246,23 @@ async def test_get_provider_open_router_missing_api_key():
assert "openrouter.ai" in exc_info.value.detail
+async def test_get_provider_deepseek_missing_api_key():
+ """DeepSeek with empty API key raises HTTPException 503."""
+ with patch("api.dependencies.get_settings") as mock_settings:
+ mock_settings.return_value = _make_mock_settings(
+ provider_type="deepseek",
+ deepseek_api_key="",
+ )
+
+ with pytest.raises(HTTPException) as exc_info:
+ get_provider()
+
+ assert exc_info.value.status_code == 503
+ assert "DEEPSEEK_API_KEY" in exc_info.value.detail
+ assert "platform.deepseek.com" in exc_info.value.detail
+
+
@pytest.mark.asyncio
async def test_get_provider_unknown_type():
"""Test that unknown provider_type raises ValueError."""
diff --git a/tests/config/test_config.py b/tests/config/test_config.py
index 35189c9e..4614f84b 100644
--- a/tests/config/test_config.py
+++ b/tests/config/test_config.py
@@ -359,6 +359,7 @@ class TestPerModelMapping:
"open_router/anthropic/claude-3-opus",
"open_router/anthropic/claude-3-haiku",
),
+ ({"MODEL": "deepseek/deepseek-chat"}, "deepseek/deepseek-chat", None),
({"MODEL": "lmstudio/qwen2.5-7b"}, "lmstudio/qwen2.5-7b", None),
({"MODEL": "llamacpp/local-model"}, "llamacpp/local-model", None),
],
@@ -494,6 +495,7 @@ class TestPerModelMapping:
assert Settings.parse_provider_type("nvidia_nim/meta/llama") == "nvidia_nim"
assert Settings.parse_provider_type("open_router/deepseek/r1") == "open_router"
+ assert Settings.parse_provider_type("deepseek/deepseek-chat") == "deepseek"
assert Settings.parse_provider_type("lmstudio/qwen") == "lmstudio"
assert Settings.parse_provider_type("llamacpp/model") == "llamacpp"
@@ -502,5 +504,6 @@ class TestPerModelMapping:
from config.settings import Settings
assert Settings.parse_model_name("nvidia_nim/meta/llama") == "meta/llama"
+ assert Settings.parse_model_name("deepseek/deepseek-chat") == "deepseek-chat"
assert Settings.parse_model_name("lmstudio/qwen") == "qwen"
assert Settings.parse_model_name("llamacpp/model") == "model"
diff --git a/tests/providers/test_deepseek.py b/tests/providers/test_deepseek.py
new file mode 100644
index 00000000..965494af
--- /dev/null
+++ b/tests/providers/test_deepseek.py
@@ -0,0 +1,181 @@
+"""Tests for DeepSeek provider."""
+
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import pytest
+
+from providers.base import ProviderConfig
+from providers.deepseek import DEEPSEEK_BASE_URL, DeepSeekProvider
+
+
+class MockMessage:
+ def __init__(self, role, content):
+ self.role = role
+ self.content = content
+
+
+class MockBlock:
+ def __init__(self, **kwargs):
+ for key, value in kwargs.items():
+ setattr(self, key, value)
+
+
+class MockRequest:
+ def __init__(self, **kwargs):
+ self.model = "deepseek-chat"
+ self.messages = [MockMessage("user", "Hello")]
+ self.max_tokens = 100
+ self.temperature = 0.5
+ self.top_p = 0.9
+ self.system = "System prompt"
+ self.stop_sequences = None
+ self.tools = []
+ self.extra_body = {}
+ self.thinking = MagicMock()
+ self.thinking.enabled = True
+ for key, value in kwargs.items():
+ setattr(self, key, value)
+
+
+def deepseek_config():
+ return ProviderConfig(
+ api_key="test_deepseek_key",
+ base_url=DEEPSEEK_BASE_URL,
+ rate_limit=10,
+ rate_window=60,
+ enable_thinking=True,
+ )
+
+
[email protected](autouse=True)
+def mock_rate_limiter():
+ """Mock the global rate limiter to prevent waiting."""
+ with patch("providers.openai_compat.GlobalRateLimiter") as mock:
+ instance = mock.get_instance.return_value
+ instance.wait_if_blocked = AsyncMock(return_value=False)
+
+ async def _passthrough(fn, *args, **kwargs):
+ return await fn(*args, **kwargs)
+
+ instance.execute_with_retry = AsyncMock(side_effect=_passthrough)
+ yield instance
+
+
+def deepseek_provider(deepseek_config):
+ return DeepSeekProvider(deepseek_config)
+
+
+def test_init(deepseek_config):
+ """Test provider initialization."""
+ with patch("providers.openai_compat.AsyncOpenAI") as mock_openai:
+ provider = DeepSeekProvider(deepseek_config)
+ assert provider._api_key == "test_deepseek_key"
+ assert provider._base_url == DEEPSEEK_BASE_URL
+ mock_openai.assert_called_once()
+
+
+def test_build_request_body_enables_thinking_for_chat_model(deepseek_provider):
+ """Thinking-enabled requests add DeepSeek's thinking payload for chat model."""
+ req = MockRequest(model="deepseek-chat")
+ body = deepseek_provider._build_request_body(req)
+
+ assert body["model"] == "deepseek-chat"
+ assert body["extra_body"]["thinking"] == {"type": "enabled"}
+ assert body["messages"][0]["role"] == "system"
+
+
+def test_build_request_body_global_disable_blocks_request_thinking():
+ """Global disable suppresses provider-side thinking even if the request enables it."""
+ provider = DeepSeekProvider(
+ ProviderConfig(
+ api_key="test_deepseek_key",
+ base_url=DEEPSEEK_BASE_URL,
+ rate_limit=10,
+ rate_window=60,
+ enable_thinking=False,
+ )
+ )
+ req = MockRequest(model="deepseek-chat")
+ body = provider._build_request_body(req)
+
+ assert "extra_body" not in body or "thinking" not in body["extra_body"]
+
+
+def test_build_request_body_request_disable_blocks_global_thinking(deepseek_provider):
+ """Request-level disable suppresses provider-side thinking when global is enabled."""
+ req = MockRequest(model="deepseek-chat")
+ req.thinking.enabled = False
+ body = deepseek_provider._build_request_body(req)
+
+ assert "extra_body" not in body or "thinking" not in body["extra_body"]
+
+
+def test_build_request_body_reasoner_skips_thinking_extra(deepseek_provider):
+ """deepseek-reasoner does not need an extra thinking payload."""
+ req = MockRequest(model="deepseek-reasoner")
+ body = deepseek_provider._build_request_body(req)
+
+ assert body["model"] == "deepseek-reasoner"
+ assert "extra_body" not in body or "thinking" not in body["extra_body"]
+
+
+def test_build_request_body_preserves_caller_thinking_override(deepseek_provider):
+ """Caller-provided thinking payload should not be overwritten."""
+ req = MockRequest(
+ model="deepseek-chat",
+ extra_body={"thinking": {"type": "manual"}},
+ )
+ body = deepseek_provider._build_request_body(req)
+
+ assert body["extra_body"]["thinking"] == {"type": "manual"}
+
+
+def test_build_request_body_preserves_reasoning_content(deepseek_provider):
+ """Thinking blocks are mirrored into reasoning_content for continuation."""
+ req = MockRequest(
+ system=None,
+ messages=[
+ MockMessage(
+ "assistant",
+ [
+ MockBlock(type="thinking", thinking="First think"),
+ MockBlock(type="text", text="Then answer"),
+ ],
+ )
+ ],
+ )
+
+ body = deepseek_provider._build_request_body(req)
+
+ assert body["messages"][0]["reasoning_content"] == "First think"
+
+
+async def test_stream_response_reasoning_content(deepseek_provider):
+ """reasoning_content deltas are emitted as thinking blocks."""
+ req = MockRequest()
+
+ mock_chunk = MagicMock()
+ mock_chunk.choices = [
+ MagicMock(
+ delta=MagicMock(content=None, reasoning_content="Thinking..."),
+ finish_reason="stop",
+ )
+ ]
+ mock_chunk.usage = MagicMock(completion_tokens=2)
+
+ async def mock_stream():
+ yield mock_chunk
+
+ with patch.object(
+ deepseek_provider._client.chat.completions, "create", new_callable=AsyncMock
+ ) as mock_create:
+ mock_create.return_value = mock_stream()
+
+ events = [event async for event in deepseek_provider.stream_response(req)]
+
+ assert any(
+ '"thinking_delta"' in event and "Thinking..." in event for event in events
+ )