---
title: "Action fastline_completion"
description: "Generates text via AI without searching the Fast Line knowledge base."
---

# Action `fastline_completion`

> **System ID:** `action_fastline_completion`

## Description

Generates an AI response from a prompt (and optionally message history) **without** searching the Fast Line knowledge base. The result is stored in a constant (`completion` or `save_response`).

## Events (scenario branches)

| Event | When it occurs |
|-------|----------------|
| *(response text)* | Success: the action returns the generated text (not a fixed branch name) |
| **error** | Missing AI prompt/key or generation error |

On success the scenario branch is **the response text itself**. For stable branching, prefer the `save_response` constant plus a separate check, or connect only the **error** branch.

## Parameters

| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| `prompt` | string | yes* | — | AI prompt. If missing — the user’s message text is used |
| `user_input` | string | no | `null` | Input text; stored in the `user_input` constant |
| `provider` | string | no | `null` | AI provider (e.g. `openai`) |
| `model` | string | no | `null` | AI model |
| `apiKey` | string | no | `null` | Access key (also `openai_token`) |
| `user_messages` | array | no | `null` | Dialog context: `[{ "role", "content" }, …]` |
| `save_response` | string | no | `"completion"` | Constant for the response |
| `customConfig` | object | no | `null` | Additional AI settings |
| `custom_fields` | object | no | `null` | Fields merged with `customConfig` |

\*In practice you need a prompt or the user’s message text, plus an AI key from config.

## Configuration example

```json
{
  "prompt": "Write a short welcome message for {{user_name}}",
  "save_response": "welcome_message"
}
```

## Related docs

- [Actions reference](/en/actionjail/reference/actions-reference.md)