MCP
Use Image Prompt Gallery with your AI agent
Connect our curated prompt library to Claude, Codex, Cursor or any MCP-compatible agent. Your agent can search real prompt cases by model, category, or meaning — free, read-only, no sign-up. Generate a key below and paste one line of config.
What your agent can do
search_promptsBrowse prompt cases with exact filters: model, category, tag, image or video.
semantic_search_promptsDescribe what you want in plain words ("warm cinematic portrait lighting") and find matching prompts.
get_promptFetch the full prompt text and preview image for one case.
list_modelsList all models with prompt collections (GPT Image 2, Seedance, Nano Banana Pro, ...).
list_categoriesList all categories and tags your agent can filter by.
Step 1 · Get your free key
One key per device, issued instantly — no account needed. The key only reads public prompt data.
Step 2 · Connect your agent
Pick your tool, copy the config, and paste it where your agent keeps MCP servers. Then just ask your agent to find you a prompt.
claude mcp add --transport http image-prompt-gallery https://imagepromptgallery.com/mcp --header "Authorization: Bearer YOUR-KEY"
Fair-use limits
| What | Per minute | Per day |
|---|---|---|
| Search, details & lists (all tools except semantic search) | 60 | 2,000 |
| semantic_search_prompts | 10 | 100 |
Your key is tied to this device and only reads public prompt data. Semantic search shares a global daily budget across all visitors, so heavy limits apply there — plain search works great for most needs.
Prefer browsing on the web? Explore the prompt library here.