---
title: "Generate AI image"
method: POST
path: "/tools/generate_ai_image"
---

# Generate AI image

`POST /tools/generate_ai_image`

Generate an image from a text prompt using your choice of AI model. Async — returns 202 with a pending job; poll GET /tool_jobs/{uid} until `status` is `completed` or `failed`.

## Request body

- object
  - `prompt` string, required — What to generate — describe the subject, style and composition
  - `model` 'flux_schnell' | 'flux_1_1_pro' | 'nano_banana' | 'gpt_image_2', required — Costs vary by model — see the credit cost shown for each
  - `aspect_ratio` '1:1' | '16:9' | '9:16' | '4:3' | '3:4' — Defaults to square
  - `reference_image_url` string, uri — Optional starting point — only some models accept one
  - `metadata` string — Arbitrary metadata string stored with the tool run

## Response `202`

Tool run accepted

- object
  - `uid` string
  - `tool` 'generate_ai_image'
  - `status` 'pending' | 'running' | 'completed' | 'failed' — pending → running → completed / failed
  - `progress` integer — 0-100 job progress, 100 once completed
  - `inputs` object — Inputs the job was created with
  - `outputs` object
    - `image_url` string, uri
  - `metadata` string, nullable — Arbitrary metadata string round-tripped from the create request
  - `self` string, uri
  - `created_at` string, date-time
  - `completed_at` string, date-time, nullable
  - `error_message` string, nullable — Human-readable failure reason (only present on failed jobs)

## Other responses

- `402` — API quota exhausted — upgrade required
- `404` — Unknown tool
- `422` — Input validation failure

---

[API](https://skmtc.dev/bannerbear/apis/bannerbear-v5-api.md) · [All operations](https://skmtc.dev/bannerbear/apis/bannerbear-v5-api/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/bannerbear/bannerbear-v5-api/revisions/6aa425420a10/schema)