# gRPC Reference
Reference of all available xAI gRPC API services. The xAI Enterprise gRPC API is a robust, high-performance gRPC interface designed for seamless integration into existing systems.
## Base URL
```
api.x.ai
```
## Authentication
For all services, authenticate with the header:
```
Authorization: Bearer <your xAI API key>
```
## Protobuf Definitions
Visit xAI API Protobuf Definitions to view and download protobuf definitions.
## Available Services
### xai_api.Chat
Chat completion service for text generation.
```protobuf
service Chat {
rpc CreateChatCompletion(ChatCompletionRequest) returns (ChatCompletionResponse);
rpc CreateChatCompletionStream(ChatCompletionRequest) returns (stream ChatCompletionChunk);
}
```
#### Example (Python)
```python
import grpc
from xai_api import chat_pb2, chat_pb2_grpc
channel = grpc.secure_channel('api.x.ai:443', grpc.ssl_channel_credentials())
stub = chat_pb2_grpc.ChatStub(channel)
metadata = [('authorization', f'Bearer {api_key}')]
request = chat_pb2.ChatCompletionRequest(
model="grok-4",
messages=[
chat_pb2.Message(role="user", content="Hello!")
]
)
response = stub.CreateChatCompletion(request, metadata=metadata)
print(response.choices[0].message.content)
```
### xai_api.Image
Image generation and editing service.
```protobuf
service Image {
rpc CreateImage(ImageRequest) returns (ImageResponse);
rpc EditImage(ImageEditRequest) returns (ImageResponse);
}
```
### xai_api.BatchMgmt
Batch management service for handling batch operations.
```protobuf
service BatchMgmt {
rpc CreateBatch(CreateBatchRequest) returns (Batch);
rpc GetBatch(GetBatchRequest) returns (Batch);
rpc ListBatches(ListBatchesRequest) returns (ListBatchesResponse);
rpc CancelBatch(CancelBatchRequest) returns (Batch);
}
```
### xai_api.Models
Model listing and information service.
```protobuf
service Models {
rpc ListModels(ListModelsRequest) returns (ListModelsResponse);
rpc GetModel(GetModelRequest) returns (Model);
rpc ListLanguageModels(ListLanguageModelsRequest) returns (ListLanguageModelsResponse);
rpc GetLanguageModel(GetLanguageModelRequest) returns (LanguageModel);
rpc ListImageGenerationModels(ListImageGenerationModelsRequest) returns (ListImageGenerationModelsResponse);
rpc GetImageGenerationModel(GetImageGenerationModelRequest) returns (ImageGenerationModel);
}
```
### xai_api.Auth
Authentication and API key management service.
```protobuf
service Auth {
rpc GetApiKey(GetApiKeyRequest) returns (ApiKey);
rpc ListApiKeys(ListApiKeysRequest) returns (ListApiKeysResponse);
rpc CreateApiKey(CreateApiKeyRequest) returns (ApiKey);
rpc DeleteApiKey(DeleteApiKeyRequest) returns (Empty);
}
```
### xai_api.Sample
Raw sampling service for advanced use cases.
```protobuf
service Sample {
rpc RawSample(RawSampleRequest) returns (RawSampleResponse);
rpc RawSampleStream(RawSampleRequest) returns (stream RawSampleChunk);
}
```
## Streaming
gRPC supports bidirectional streaming for real-time responses:
```python
# Streaming example
stream = stub.CreateChatCompletionStream(request, metadata=metadata)
for chunk in stream:
if chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="")
```
## xAI Python SDK
For Python, the xAI SDK covers all features and uses gRPC for optimal performance:
```python
from xai_sdk import Client
client = Client(api_key="your-api-key")
# SDK handles gRPC internally
response = client.chat.completions.create(
model="grok-4",
messages=[{"role": "user", "content": "Hello!"}]
)
```
The xAI SDK allows interaction with all products:
- Collections
- Voice API
- API key management
- And more
## Deprecated Services
### Messages (Anthropic Compatible)
**Status**: Deprecated
The Anthropic SDK compatibility is fully deprecated. Migrate to the Responses API or gRPC.
## Error Handling
gRPC uses standard status codes:
| Code | Description |
|------|-------------|
| OK (0) | Success |
| INVALID_ARGUMENT (3) | Invalid request |
| UNAUTHENTICATED (16) | Auth failed |
| PERMISSION_DENIED (7) | Access denied |
| NOT_FOUND (5) | Resource not found |
| RESOURCE_EXHAUSTED (8) | Rate limited |
| INTERNAL (13) | Server error |
| UNAVAILABLE (14) | Service unavailable |
## Best Practices
1. **Use connection pooling**: Reuse gRPC channels
2. **Handle streaming properly**: Process chunks as they arrive
3. **Implement retries**: Use exponential backoff for transient errors
4. **Set deadlines**: Prevent hanging requests
5. **Use the SDK**: The xAI SDK handles gRPC complexity for you
