Simple, credit-aware API
Authenticate with a personal API key and turn an idea or public image URL into a ready-to-use prompt.
Read API docsDEVELOPERS
Build with Find Image Prompt
Use the same image-prompt intelligence, account and credit balance as the web app.
Install the Skill
Give AI agents clear operation selection, key safety, request patterns, and error recovery for this API.
Download SkillSKILL.md · Agent-readyDEVELOPER GUIDE
An image to prompt Skill API for real creative workflows
Find Image Prompt gives creative tools, product teams, and AI agents one consistent way to translate a brief or reference image into model-specific visual direction. Choose a prompt operation, send text or a public image URL, and receive a ready-to-use result.
Our image to prompt API is built for creative workflows, asset libraries, prompt editors, and agent tools that need repeatable prompt quality without managing a separate model integration. It turns visual intent into a structured, model-aware instruction that a person or image model can act on.
WHAT YOU CAN BUILD
API and Skill, designed for different jobs
Use the API when your product needs a stable integration. Use the image to prompt Skill when an AI agent needs the same integration plus decision rules.
Image to Prompt API
Turn a public reference image into a faithful image-generation prompt. The service captures subject, setting, composition, lighting, materials, and the style signature instead of returning generic “high quality” filler.
Model-aware prompt writing
Create prompts for GPT Image, Midjourney, Flux, Stable Diffusion, Nano Banana, ChatGPT Images, and Gemini. Each operation adapts syntax and emphasis to the target model.
Image to Prompt Skill for agents
The downloadable Skill tells an agent which API operation to select, how to protect the API key, how to handle image URLs, and when to stop rather than retry a safety refusal.
USE CASES
Where teams use the API and Skill
Use the API inside a product or backend. Use the Skill when an agent needs to make the same choices safely on a user's behalf.
Reference-image prompt editor
A design tool accepts a public image URL, calls image_to_prompt, and gives creators an editable prompt that preserves the source image's composition, lighting, and visual style.
Campaign brief to prompt library
A marketing system turns product briefs into consistent visual directions with text_to_prompt, then saves the response, request ID, and approved prompt for future campaign variants.
Creative assistant with guardrails
An agent installs the image to prompt Skill. It selects image analysis when a user shares a reference, selects text generation for an idea, keeps the API key in an environment secret, and stops on a safety refusal.
Outcome: a predictable agent workflow without embedding custom integration logic in every conversation.DESIGN PRINCIPLES
Why the API and Skill are designed this way
One shared product boundary. The API reuses the same credit accounting, safety checks, moderation, and account policy as Find Image Prompt. Paid-only API access keeps server-side automation inside a clear commercial and abuse-control boundary.
Explicit choices over magic. The API exposes operations such as image-to-prompt, image description, and model-specific prompt generation. The Skill turns those choices into clear agent instructions, so behavior stays predictable instead of relying on an agent to infer an undocumented workflow.
Portable by default. The OpenAPI schema, downloadable SKILL.md, environment-variable authentication, and direct JSON responses make the integration easy to audit, install, and submit to skill directories.
FAQ
Common developer questions
Who can use the image to prompt Skill API?
API access is available to paid Starter and Pro accounts. API calls use the same account balance and safety limits as the website; prompt requests cost 1 credit.
Which image inputs are supported?
Use a public direct HTTPS URL to a PNG, JPEG, or WebP file up to 10 MB. The API does not accept private-network or local URLs.
How does the image to prompt Skill work?
Download the Skill file, configure the API key as an environment secret, and let the agent select an operation based on whether the user supplied text, an image, or both. It also gives the agent safety and error-handling rules.
Can I revoke access immediately?
Yes. Revoke a personal API key from this page. Requests made with that key stop working right away.
API access is for paid plansUpgrade to Starter or Pro to create API keys. This keeps server-to-server access inside the paid account safety boundary.
Sign in to upgradeAPI access requires a paid plan. Prompt requests cost 1 credit and remain subject to account-level usage and safety limits.