Promptcelerator
免費的 AI 影片提示詞資料庫,每個提示詞都經過實測,旁邊附有生成的影片,並標明創作者。用任何語言以白話描述影片:你的 AI 會找出符合需求的提示詞、展示影片,並能根據你的想法調整最佳提示詞。僅供讀取,無需帳戶或金鑰。
託管 MCP 伺服器
npx add-mcp 'https://promptcelerator.com/mcp'可安裝到 Claude Code、Codex、Cursor 等客戶端
文件
Connect Claude, ChatGPT or another assistant to the library. Tell it what you want to make: it finds the prompts that fit, shows you the videos and adapts the best one to your idea.
The address to give your AI
https://promptcelerator.com/mcp
Free, with no account and no key. Your AI can only read the library, it cannot change anything.
Then just ask
- I want a slow-motion perfume ad. Which prompts should I start from?
- Find me a documentary-style prompt for Seedance and adapt it to a fishing village.
- What works for an anime fight scene in vertical format?
Connect it in a minute
Claude
On the web, desktop or mobile app: in Settings, open Connectors, add a custom connector and paste the address.
Claude Code
claude mcp add --transport http promptcelerator https://promptcelerator.com/mcp
ChatGPT
On a paid plan: in Settings, open Apps & Connectors, then Advanced settings, and turn on developer mode. Go back, choose Create, paste the address and select no authentication.
Cursor
VS Code
Codex
codex mcp add promptcelerator --url https://promptcelerator.com/mcp
Gemini CLI
gemini mcp add --transport http promptcelerator https://promptcelerator.com/mcp
Another assistant
Any assistant that supports MCP over HTTP can use the address above. Many of them read a configuration like this one:
{
"mcpServers": {
"promptcelerator": {
"url": "https://promptcelerator.com/mcp"
}
}
}
Without MCP
The same search is available as a simple address, free and without a key. Describe the video in plain words, in any language:
curl "https://promptcelerator.com/api/v1/search?q=a+slow+motion+perfume+ad&limit=3"
/openapi.json: to import the search into a custom GPT or any assistant that reads OpenAPI/llms.txt: the summary of the site for language models
Your AI is asked to name the creator and link to the original post whenever it shows or adapts a prompt: the prompts and videos belong to them.