diff --git a/docs/my-website/docs/providers/openclaw.md b/docs/my-website/docs/providers/openclaw.md
new file mode 100644
index 00000000000..7f2cb418110
--- /dev/null
+++ b/docs/my-website/docs/providers/openclaw.md
@@ -0,0 +1,186 @@
+import Tabs from '@theme/Tabs';
+import TabItem from '@theme/TabItem';
+
+# OpenClaw
+
+[OpenClaw](https://openclaw.ai) is an AI agent framework that exposes an OpenAI-compatible HTTP endpoint. It allows you to interact with AI agents that have access to tools, memory, and custom configurations.
+
+## Key Features
+
+- **Agent Targeting**: Route requests to specific agents via the model field (`openclaw/main`, `openclaw/research`)
+- **Session Persistence**: Maintain conversation context across requests using the `user` field
+- **Full Tool Access**: Agents can execute code, browse the web, manage files, and more
+- **Streaming Support**: Real-time SSE streaming responses
+
+## Quick Start
+
+```python
+import litellm
+
+response = litellm.completion(
+    model="openclaw/main",  # Target the 'main' agent
+    api_base="http://localhost:18789",
+    api_key="your-gateway-token",
+    messages=[{"role": "user", "content": "Hello!"}]
+)
+print(response.choices[0].message.content)
+```
+
+## Environment Variables
+
+```bash
+export OPENCLAW_API_BASE="http://localhost:18789"  # Gateway URL
+export OPENCLAW_API_KEY="your-gateway-token"       # Auth token
+```
+
+## Usage
+
+### SDK Usage
+
+<Tabs>
+<TabItem value="sdk" label="SDK">
+
+```python
+import litellm
+
+# Basic completion
+response = litellm.completion(
+    model="openclaw/main",
+    messages=[{"role": "user", "content": "What can you do?"}]
+)
+
+# Streaming
+response = litellm.completion(
+    model="openclaw/main",
+    messages=[{"role": "user", "content": "Tell me a story"}],
+    stream=True
+)
+for chunk in response:
+    print(chunk.choices[0].delta.content or "", end="")
+
+# With session persistence (same user = same conversation)
+response = litellm.completion(
+    model="openclaw/main",
+    messages=[{"role": "user", "content": "Remember my name is Alice"}],
+    user="alice-session-123"
+)
+```
+
+</TabItem>
+<TabItem value="proxy" label="LiteLLM Proxy">
+
+1. Add to your `config.yaml`:
+
+```yaml
+model_list:
+  - model_name: openclaw-main
+    litellm_params:
+      model: openclaw/main
+      api_base: http://localhost:18789
+      api_key: os.environ/OPENCLAW_API_KEY
+  
+  - model_name: openclaw-research
+    litellm_params:
+      model: openclaw/research
+      api_base: http://localhost:18789
+      api_key: os.environ/OPENCLAW_API_KEY
+```
+
+2. Start the proxy:
+
+```bash
+litellm --config config.yaml
+```
+
+3. Make requests:
+
+```bash
+curl http://localhost:4000/v1/chat/completions \
+  -H "Content-Type: application/json" \
+  -H "Authorization: Bearer sk-1234" \
+  -d '{
+    "model": "openclaw-main",
+    "messages": [{"role": "user", "content": "Hello!"}]
+  }'
+```
+
+</TabItem>
+</Tabs>
+
+## Targeting Different Agents
+
+OpenClaw can run multiple agents with different configurations. Target them via the model field:
+
+```python
+# Main agent (default)
+litellm.completion(model="openclaw/main", ...)
+
+# Research agent with web search tools
+litellm.completion(model="openclaw/research", ...)
+
+# Custom agent
+litellm.completion(model="openclaw/my-custom-agent", ...)
+```
+
+## Supported Parameters
+
+| Parameter | Type | Description |
+|-----------|------|-------------|
+| `model` | string | Agent to target: `openclaw/<agent-id>` |
+| `messages` | array | Conversation messages |
+| `stream` | boolean | Enable SSE streaming |
+| `temperature` | float | Sampling temperature |
+| `max_tokens` | integer | Maximum tokens to generate |
+| `user` | string | Session key for conversation persistence |
+| `tools` | array | Tool definitions (agents have built-in tools) |
+| `tool_choice` | string/object | Tool selection preference |
+
+## OpenClaw Setup
+
+To use OpenClaw with LiteLLM:
+
+1. Install and start OpenClaw:
+```bash
+npm install -g openclaw
+openclaw gateway
+```
+
+2. Enable the HTTP endpoint in your OpenClaw config:
+```json
+{
+  "gateway": {
+    "http": {
+      "endpoints": {
+        "chatCompletions": { "enabled": true }
+      }
+    }
+  }
+}
+```
+
+3. Configure authentication (optional but recommended):
+```bash
+export OPENCLAW_GATEWAY_TOKEN="your-secure-token"
+```
+
+For more details, see the [OpenClaw documentation](https://docs.openclaw.ai).
+
+## Troubleshooting
+
+### Connection Refused
+Ensure the OpenClaw gateway is running and the API base URL is correct:
+```bash
+curl http://localhost:18789/health
+```
+
+### Authentication Failed
+Verify your gateway token matches the one configured in OpenClaw:
+```bash
+openclaw gateway status
+```
+
+### Agent Not Found
+Check available agents:
+```bash
+openclaw agents list
+```
diff --git a/litellm/__init__.py b/litellm/__init__.py
index a74a79635f0..56e6eb0c614 100644
--- a/litellm/__init__.py
+++ b/litellm/__init__.py
@@ -1426,6 +1426,7 @@ if TYPE_CHECKING:
     from .llms.llamafile.chat.transformation import LlamafileChatConfig as _LlamafileChatConfig
     from .llms.lm_studio.chat.transformation import LMStudioChatConfig as _LMStudioChatConfig
     from .llms.lm_studio.embed.transformation import LmStudioEmbeddingConfig as _LmStudioEmbeddingConfig
+    from .llms.openclaw.chat.transformation import OpenClawChatConfig as _OpenClawChatConfig
     from .llms.watsonx.embed.transformation import IBMWatsonXEmbeddingConfig as _IBMWatsonXEmbeddingConfig
     from .llms.vertex_ai.gemini.vertex_and_google_ai_studio_gemini import VertexGeminiConfig as _VertexGeminiConfig
 
@@ -1444,6 +1445,7 @@ if TYPE_CHECKING:
     LlamafileChatConfig: Type[_LlamafileChatConfig]
     LMStudioChatConfig: Type[_LMStudioChatConfig]
     LmStudioEmbeddingConfig: Type[_LmStudioEmbeddingConfig]
+    OpenClawChatConfig: Type[_OpenClawChatConfig]
     IBMWatsonXEmbeddingConfig: Type[_IBMWatsonXEmbeddingConfig]
     VertexAIConfig: Type[_VertexGeminiConfig]  # Alias for VertexGeminiConfig
 
diff --git a/litellm/_lazy_imports_registry.py b/litellm/_lazy_imports_registry.py
index 0e52e9a59eb..126f517089a 100644
--- a/litellm/_lazy_imports_registry.py
+++ b/litellm/_lazy_imports_registry.py
@@ -1022,6 +1022,7 @@ _LLM_CONFIGS_IMPORT_MAP = {
     "VLLMConfig": (".llms.vllm.completion.transformation", "VLLMConfig"),
     "DeepSeekChatConfig": (".llms.deepseek.chat.transformation", "DeepSeekChatConfig"),
     "LMStudioChatConfig": (".llms.lm_studio.chat.transformation", "LMStudioChatConfig"),
+    "OpenClawChatConfig": (".llms.openclaw.chat.transformation", "OpenClawChatConfig"),
     "LmStudioEmbeddingConfig": (
         ".llms.lm_studio.embed.transformation",
         "LmStudioEmbeddingConfig",
diff --git a/litellm/constants.py b/litellm/constants.py
index 3c84547d7ce..1435731041a 100644
--- a/litellm/constants.py
+++ b/litellm/constants.py
@@ -415,6 +415,7 @@ LITELLM_CHAT_PROVIDERS = [
     "hosted_vllm",
     "llamafile",
     "lm_studio",
+    "openclaw",
     "galadriel",
     "gradient_ai",
     "github_copilot",  # GitHub Copilot Chat API
@@ -619,6 +620,7 @@ openai_compatible_providers: List = [
     "hosted_vllm",
     "llamafile",
     "lm_studio",
+    "openclaw",
     "galadriel",
     "github_copilot",  # GitHub Copilot Chat API
     "chatgpt",  # ChatGPT subscription API
diff --git a/litellm/llms/openclaw/chat/transformation.py b/litellm/llms/openclaw/chat/transformation.py
new file mode 100644
index 00000000000..e0f442d72a5
--- /dev/null
+++ b/litellm/llms/openclaw/chat/transformation.py
@@ -0,0 +1,83 @@
+"""
+Translate from OpenAI's `/v1/chat/completions` to OpenClaw's `/v1/chat/completions`
+
+OpenClaw is an AI agent framework that exposes an OpenAI-compatible HTTP endpoint.
+https://docs.openclaw.ai
+
+Key features:
+- Target specific agents via model field: `openclaw/main`, `openclaw/research`
+- Session persistence via `user` field
+- Streaming support (SSE)
+"""
+
+from typing import List, Optional, Tuple
+
+from litellm.secret_managers.main import get_secret_str
+
+from ...openai.chat.gpt_transformation import OpenAIGPTConfig
+
+
+class OpenClawChatConfig(OpenAIGPTConfig):
+    """
+    OpenClaw configuration for chat completions.
+    
+    OpenClaw agents are targeted via the model field:
+    - `openclaw/main` -> main agent
+    - `openclaw/research` -> research agent
+    - `openclaw/<agent-id>` -> any configured agent
+    
+    Environment variables:
+    - OPENCLAW_API_BASE: Gateway URL (e.g., http://localhost:18789)
+    - OPENCLAW_API_KEY: Gateway auth token
+    """
+
+    def get_supported_openai_params(self, model: str) -> List[str]:
+        """OpenClaw supports standard OpenAI params plus user for session persistence."""
+        return [
+            "stream",
+            "stop",
+            "temperature",
+            "top_p",
+            "max_tokens",
+            "max_completion_tokens",
+            "presence_penalty",
+            "frequency_penalty",
+            "logit_bias",
+            "user",  # Used for session key derivation
+            "n",
+            "tools",
+            "tool_choice",
+            "response_format",
+        ]
+
+    def _get_openai_compatible_provider_info(
+        self, api_base: Optional[str], api_key: Optional[str]
+    ) -> Tuple[Optional[str], Optional[str]]:
+        """
+        Get OpenClaw API base and key from environment or parameters.
+        
+        OpenClaw requires an auth token when gateway.auth.mode is set.
+        """
+        api_base = api_base or get_secret_str("OPENCLAW_API_BASE")
+        dynamic_api_key = api_key or get_secret_str("OPENCLAW_API_KEY") or ""
+        return api_base, dynamic_api_key
+
+    def map_openai_params(
+        self,
+        non_default_params: dict,
+        optional_params: dict,
+        model: str,
+        drop_params: bool,
+    ) -> dict:
+        """
+        Map OpenAI params to OpenClaw format.
+        
+        OpenClaw is fully OpenAI-compatible, so minimal transformation needed.
+        The model field can include agent targeting: openclaw/main -> agent:main
+        """
+        return super().map_openai_params(
+            non_default_params=non_default_params,
+            optional_params=optional_params,
+            model=model,
+            drop_params=drop_params,
+        )
diff --git a/litellm/main.py b/litellm/main.py
index 13361c644cb..27192781ae7 100644
--- a/litellm/main.py
+++ b/litellm/main.py
@@ -4813,6 +4813,7 @@ def embedding(  # noqa: PLR0915
             custom_llm_provider == "openai_like"
             or custom_llm_provider == "llamafile"
             or custom_llm_provider == "lm_studio"
+            or custom_llm_provider == "openclaw"
         ):
             api_base = (
                 api_base or litellm.api_base or get_secret_str("OPENAI_LIKE_API_BASE")
diff --git a/litellm/types/utils.py b/litellm/types/utils.py
index 6c330d0f83c..4e5e7200513 100644
--- a/litellm/types/utils.py
+++ b/litellm/types/utils.py
@@ -3107,6 +3107,7 @@ class LlmProviders(str, Enum):
     POE = "poe"
     CHUTES = "chutes"
     XIAOMI_MIMO = "xiaomi_mimo"
+    OPENCLAW = "openclaw"
 
 
 # Create a set of all provider values for quick lookup
diff --git a/litellm/utils.py b/litellm/utils.py
index 7c4eec7ba32..42efab6b5ff 100644
--- a/litellm/utils.py
+++ b/litellm/utils.py
@@ -7832,6 +7832,7 @@ class ProviderConfigManager:
             LlmProviders.HOSTED_VLLM: (lambda: litellm.HostedVLLMChatConfig(), False),
             LlmProviders.LLAMAFILE: (lambda: litellm.LlamafileChatConfig(), False),
             LlmProviders.LM_STUDIO: (lambda: litellm.LMStudioChatConfig(), False),
+            LlmProviders.OPENCLAW: (lambda: litellm.OpenClawChatConfig(), False),
             LlmProviders.GALADRIEL: (lambda: litellm.GaladrielChatConfig(), False),
             LlmProviders.REPLICATE: (lambda: litellm.ReplicateConfig(), False),
             LlmProviders.HUGGINGFACE: (lambda: litellm.HuggingFaceChatConfig(), False),