> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bidsmith.pro/llms.txt
> Use this file to discover all available pages before exploring further.

# Generate

> Run model inference on the ARIS network.

Send a prompt to an eligible worker node and receive generated output plus usage metadata.

## Request body

<ParamField name="prompt" body="prompt" type="string" required>
  Input text for inference.
</ParamField>

<ParamField name="model" body="model" type="string">
  Model identifier (for example `tinyllama`).
</ParamField>

<ParamField name="max_tokens" body="max_tokens" type="integer">
  Maximum output tokens. Default is `256`.
</ParamField>

<ParamField name="temperature" body="temperature" type="number">
  Sampling temperature between `0.0` and `2.0`. Default is `0.7`.
</ParamField>

## Example request

```json theme={null}
{
  "prompt": "Explain quantum tunneling in one paragraph.",
  "model": "tinyllama",
  "max_tokens": 256,
  "temperature": 0.7
}
```

## Success response

<ResponseField name="id" type="string">
  Unique job ID for this generation request.
</ResponseField>

<ResponseField name="node" type="string">
  Worker node that executed the inference.
</ResponseField>

<ResponseField name="output" type="string">
  Generated completion text.
</ResponseField>

<ResponseField name="usage" type="object">
  Usage object with token and credits metadata.
</ResponseField>

## Response example

```json theme={null}
{
  "id": "job_abc123",
  "node": "node-9006",
  "usage": {
    "tokens": 142,
    "credits_deducted": 1.42
  },
  "output": "Quantum tunneling occurs when..."
}
```

## Error responses

| Status | Meaning                     | Retry |
| ------ | --------------------------- | ----- |
| `400`  | Invalid request parameters. | No    |
| `401`  | Missing or invalid API key. | No    |
| `402`  | Insufficient credits.       | No    |
| `429`  | Rate limited.               | Yes   |
| `503`  | No healthy nodes available. | Yes   |