find-image-prompt

---
name: find-image-prompt
description: Create production-ready AI image prompts and image descriptions with the Find Image Prompt image-to-prompt API.
metadata:
  author: Find Image Prompt
  version: 1.2.0
  homepage: https://findimageprompt.com/
  documentation: https://findimageprompt.com/dev/docs
  keywords: image to prompt, image to prompt skill, image to prompt API, AI image prompts
---

# Find Image Prompt Skill

Use this Skill when a user wants to turn an idea or a publicly accessible reference image into a prompt for an image model.

Find Image Prompt homepage: [https://findimageprompt.com/](https://findimageprompt.com/). For API documentation, examples, and API-key management, visit [Find Image Prompt Developers](https://findimageprompt.com/dev).

## Authentication

For the Find Image Prompt API, require `FIND_IMAGE_PROMPT_API_KEY` from a paid Starter or Pro account. The bundled DeepSeek adapter can alternatively use `OPENROUTER_API_KEY`, `OPENAI_API_KEY`, or `GEMINI_API_KEY` to generate similar prompt text directly with the user's chosen model provider. Never ask a user to paste any key into chat, include a key in URLs, logs, output, or source control.

## Endpoint

`POST https://findimageprompt.com/api/v1/prompts`

Use JSON. Successful replies include `result`, `creditsUsed`, and `requestId`.

## Choose the operation

- `text_to_prompt`: turn a text idea into a reusable prompt. Requires `text`.
- `image_to_prompt`: recreate a public reference image as a prompt. Requires `image_url`.
- `describe_image`: describe visible content in a public image. Requires `image_url`; optionally send `description_option` (`detail`, `brief`, `objects`, `artStyle`, `extractText`, or `custom` plus `question`).
- `chatgpt_image_prompt`: make a ChatGPT Images-ready prompt from text, image, or both.
- `gemini_photo_prompt`: make a Gemini photo-generation prompt from text, image, or both.

For a model-specific prompt, use `target_model`: `Nano Banana`, `GPT Image`, `Flux`, `Midjourney`, or `Stable Diffusion`. Use `prompt_format` only for prompt operations: `General prompt`, `Detailed prompt`, `Structured prompt`, `Graphic design`, or `JSON`.

## Request example

```bash
curl https://findimageprompt.com/api/v1/prompts \
  -H "Authorization: Bearer $FIND_IMAGE_PROMPT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"operation":"text_to_prompt","text":"Rainy Kyoto bookshop at blue hour","target_model":"Midjourney"}'
```

## Safety and errors

Only send user-authorized content. `image_url` must be a direct public HTTPS PNG, JPEG, or WebP file under 10 MB. Do not retry `400` safety refusals. Ask the user to update the request. For `401`, ask the operator to configure a valid key; do not guess one. For `402`, explain that the account needs credits. For `403`, explain that API access requires a paid Starter or Pro plan. For `429` or `5xx`, retry once with exponential backoff, then return the error and `requestId`.

## Response handling

Return `result` exactly as the generated prompt or description. Mention credits only when the user asks or an error requires it. Do not claim an image was generated: this API creates prompt text and descriptions.

## Official links

- Homepage: https://findimageprompt.com/
- Developers and API keys: https://findimageprompt.com/dev
- API documentation: https://findimageprompt.com/dev/docs
- OpenAPI schema: https://findimageprompt.com/api/openapi.json