Chrome extension · free · local-first

Turn great references into your next creation.

PromptDirector keeps the work and the creative clues behind it. Open a work, post or article, capture it, and the image, the video, the original prompt and the source page land in a private library on your own machine — ready to find, reuse and hand to your AI agent.

Add to Chrome — free Browse curated cases How it works

No account
Local storage
No analytics or ads
Open source
Agent-ready (MCP)

PromptDirector library: navigation on the left, search on top and a wall of captured image and video references — One local library for images, videos, prompts, sources and notes.

What it does

Most prompt tools start from an empty box. The references you actually want are already out there — in a gallery post, a tweet, a video breakdown. PromptDirector works where you find them.

Capture the work and its clues

One click keeps the image or video together with the publicly available prompt, the source page, the author and the capture time — not just a screenshot.

A library that stays yours

Cases, media, tags, Creative Skills and settings live in your browser's local storage. No account, no analytics, no developer-operated data server.

Find it again

Browse visually, filter by content type and tags, sort by project, and keep multi-image works, article structure and table layouts intact.

Put it to work

Extract reusable Creative Skills, compose new images or video through your own AI service, and let a local MCP agent search, read and save into the same library.

Where it works

Midjourney
Jimeng
Higgsfield
LiblibAI
Krea
X / Twitter
Xiaohongshu
Weibo
Reddit
Pinterest
Behance
ArtStation
Feishu documents
WeChat articles
HeyBox

Plus general webpage capture, text and image selection, screenshots, and local imports — images, videos, audio, DOCX, PDF, Markdown, HTML and RTF, including DOCX text, tables and embedded images. Multi-case posts can be split into individual image-and-prompt groups, and a whole article can be saved with its structure.

How it works

Open the work, post or article Open the page in Chrome. PromptDirector reads the tab only after you start a capture.
Capture it Use the side panel, the right-click entry or the keyboard shortcut, then correct the detected subject, text and media if needed.
Preview and save Confirming the subject does not confirm its media: only the media you explicitly approve is written to a case.

A closer look

The screenshots below come from PromptDirector 1.26.0 with the interface in English; only the composer image still shows an earlier build. PromptDirector follows your Chrome language, so your own copy matches your browser.

PromptDirector library with a wall of captured image and video cases — The library — cases, projects and filters in one visual wall.

Capture preview confirming the subject, text and media before saving — Capture — review the detected subject, prompt and media before anything is saved.

Composer using saved cases as references for a new generation — The composer — turn saved references into the next image or video.

Before you install

Chrome 134 or newer.
Capture and organizing need no AI service and no account.
Capture works on pages you can already access; it does not bypass logins, paywalls, access controls or CAPTCHAs.
Original prompts must be publicly visible on the page. Login state, lazy loading and site changes can affect results.
Preview before saving — public URLs may be shown for review, but only media you confirm is written to a case.
Optional AI features use your own API key and run only when you start them.
Website access is optional and requested per feature; it is not granted at install time and does not monitor browsing.
Your cases stay on your machine. See the privacy policy for the data each action handles.

Free to use. If you want your reference library to be structured data instead of a folder of screenshots, install it and save your first case — it takes about five minutes.