> ## Documentation Index
> Fetch the complete documentation index at: https://docs.chucky.cloud/llms.txt
> Use this file to discover all available pages before exploring further.
# Channeling API
> Anthropic-compatible API with Chucky JWT authentication
The Channeling API is a drop-in replacement for the Anthropic Messages API. Use your existing Anthropic SDK or HTTP calls, but authenticate with Chucky JWT tokens instead of Anthropic API keys.
## Endpoint
```
https://channeling.chucky.cloud/v1/messages
```
***
## Authentication
Replace the Anthropic API key with your Chucky JWT token in the `x-api-key` header:
```bash theme={null}
curl -X POST https://channeling.chucky.cloud/v1/messages \
-H "x-api-key: YOUR_CHUCKY_JWT" \
-H "anthropic-version: 2023-06-01" \
-H "content-type: application/json" \
-d '{
"model": "claude-sonnet-4-5-20250929",
"max_tokens": 1024,
"messages": [
{"role": "user", "content": "Hello, Claude!"}
]
}'
```
### Creating Tokens
Tokens are created server-side with your HMAC secret:
```typescript theme={null}
import { createToken, createBudget } from '@chucky.cloud/sdk';
const token = await createToken({
userId: 'user-123',
projectId: process.env.CHUCKY_PROJECT_ID,
secret: process.env.CHUCKY_HMAC_SECRET,
budget: createBudget({
aiDollars: 5.00,
computeHours: 1,
window: 'day',
}),
});
```
See [Authentication](/authentication) for full token creation guide.
***
## Supported Endpoints
| Endpoint | Method | Description |
| -------------- | ------ | ----------------------------------------- |
| `/v1/messages` | POST | Create a message (Anthropic Messages API) |
### Request Format
The request body follows the [Anthropic Messages API](https://docs.anthropic.com/en/api/messages) format exactly:
```json theme={null}
{
"model": "claude-sonnet-4-5-20250929",
"max_tokens": 1024,
"messages": [
{"role": "user", "content": "Explain quantum computing"}
],
"system": "You are a helpful assistant.",
"stream": false
}
```
### Response Format
Responses match the Anthropic API format:
```json theme={null}
{
"id": "msg_01XFDUDYJgAACzvnptvVoYEL",
"type": "message",
"role": "assistant",
"content": [
{
"type": "text",
"text": "Quantum computing is..."
}
],
"model": "claude-sonnet-4-5-20250929",
"stop_reason": "end_turn",
"usage": {
"input_tokens": 25,
"output_tokens": 150
}
}
```
***
## Streaming
Enable streaming by setting `"stream": true`:
```bash theme={null}
curl -X POST https://channeling.chucky.cloud/v1/messages \
-H "x-api-key: YOUR_CHUCKY_JWT" \
-H "anthropic-version: 2023-06-01" \
-H "content-type: application/json" \
-d '{
"model": "claude-sonnet-4-5-20250929",
"max_tokens": 1024,
"messages": [{"role": "user", "content": "Write a poem"}],
"stream": true
}'
```
Streaming uses Server-Sent Events (SSE) matching the Anthropic format.
***
## Multi-Provider Support
The Channeling API supports multiple AI providers through model routing:
### Anthropic (Default)
```json theme={null}
{"model": "claude-sonnet-4-5-20250929"}
{"model": "claude-opus-4-5-20251101"}
{"model": "claude-3-5-haiku-20241022"}
```
### OpenAI
```json theme={null}
{"model": "gpt-5.2"}
{"model": "gpt-5"}
{"model": "o4-mini"}
```
Requires OpenAI API key configured in your project settings.
### OpenRouter
Use the `or:` prefix for explicit routing, or any unknown model name:
```json theme={null}
{"model": "or:mistralai/mistral-large"}
{"model": "or:deepseek/deepseek-chat"}
{"model": "or:mistralai/devstral-2512:free"}
```
Requires OpenRouter API key configured in your project settings.
See [Models & Providers](/concepts/models) for full provider documentation.
***
## Using with Anthropic SDKs
### TypeScript/JavaScript
```typescript theme={null}
import Anthropic from '@anthropic-ai/sdk';
const client = new Anthropic({
apiKey: CHUCKY_JWT_TOKEN,
baseURL: 'https://channeling.chucky.cloud',
});
const message = await client.messages.create({
model: 'claude-sonnet-4-5-20250929',
max_tokens: 1024,
messages: [{ role: 'user', content: 'Hello!' }],
});
```
### Python
```python theme={null}
import anthropic
client = anthropic.Anthropic(
api_key=CHUCKY_JWT_TOKEN,
base_url='https://channeling.chucky.cloud',
)
message = client.messages.create(
model='claude-sonnet-4-5-20250929',
max_tokens=1024,
messages=[{'role': 'user', 'content': 'Hello!'}],
)
```
***
## Budget Enforcement
The Channeling API enforces budgets embedded in JWT tokens:
* **AI Budget**: Limits total AI spend per user
* **Compute Budget**: Not applicable (no sandbox execution)
When budget is exceeded, requests return a `429` error:
```json theme={null}
{
"error": {
"type": "rate_limit_error",
"message": "AI budget exceeded for user"
}
}
```
***
## Cost Tracking
Every request is tracked against the user's budget. Costs are calculated based on:
1. Model used
2. Input/output tokens
3. [Custom pricing](/billing#custom-model-pricing) (if configured)
View usage in the [dashboard](https://app.chucky.cloud).
***
## Use Cases
### Direct API Access
Use HTTP calls without the Chucky SDK:
```bash theme={null}
# From any language or tool that can make HTTP requests
curl -X POST https://channeling.chucky.cloud/v1/messages ...
```
### Existing Anthropic Integrations
Point existing Anthropic SDK code to Chucky by changing the base URL:
```typescript theme={null}
// Before
const client = new Anthropic({ apiKey: ANTHROPIC_API_KEY });
// After
const client = new Anthropic({
apiKey: CHUCKY_JWT_TOKEN,
baseURL: 'https://channeling.chucky.cloud',
});
```
### Third-Party Tools
Any tool that supports custom Anthropic endpoints can use Channeling:
* VS Code extensions
* CLI tools
* Automation platforms
***
## Error Handling
| Status | Error Type | Description |
| ------ | ----------------------- | ----------------------- |
| 400 | `invalid_request_error` | Malformed request body |
| 401 | `authentication_error` | Invalid or expired JWT |
| 429 | `rate_limit_error` | Budget exceeded |
| 500 | `api_error` | Upstream provider error |
***
## Next Steps
Create JWT tokens with budgets
Configure multi-provider support
Understand costs and pricing
View usage and manage projects