adracle

adracle

The Facebook Ad Library, captured to your disk.

A Chrome side-panel extension that crawls an Ad Library query and saves everything locally — creative text, landing URLs, original video and image files, EU transparency data. No cloud, no account, no telemetry. And the export isn’t a spreadsheet to scroll: it’s a corpus, shaped to be read by an AI.

Download adracle‑chrome.zip How to install

172 KB · Chrome (Manifest V3) · installs unpacked — not on the Web Store, by design · latest build, straight from the release

The adracle side panel: a capture form and a list of crawl jobs in the dark Chamfer Glass theme

Read this before installing

adracle automates browsing of the Ad Library, which violates Meta’s Terms of Service no matter how gently it crawls. The Facebook account you crawl with may be flagged or restricted. The extension is disabled until you accept this risk in its settings — and that is also why it isn’t distributed through the Chrome Web Store: it wouldn’t pass review, and it shouldn’t pretend otherwise.

Everything stays on your machine. The only network requests adracle makes are the ones Facebook’s own page would make.

What it captures

3 layers per ad
metadata

Ads & snapshots

One record per Library ID: creative text, headline, CTA, decoded landing URLs, platforms, dates, status — plus a snapshot per crawl, so re-crawling the same page builds a change history. Carousel and DCO variants are captured card by card.

assets

Original files

The actual mp4s and images from Facebook’s CDN — highest-quality renditions, saved under AdLibrary/media/ in your browser’s download folder and referenced by local path in the export. No screenshots, no re-encodes.

eu dsa

EU transparency

The DSA disclosure block per ad: total EU reach, the country × age × gender reach grid, targeting criteria, payer and beneficiary. Fetched in a gentle second pass after the feed crawl.

Built to be read by AI

the best use case so far

Two hundred captured ads are not something you browse — they’re something you interrogate. The export schema is designed for model consumption, not human eyes: one compact JSON file, one record per ad with its capture history merged in, advertiser pages deduplicated, DCO variants resolved card by card, media referenced by local file path, empty fields omitted. A whole crawl fits in one context window.

Drop the file into Claude, ChatGPT, or a coding agent and ask about the strategy behind the ads — the patterns no amount of scrolling the Ad Library would surface. The original media files sit next to the JSON, so a multimodal model can look at the creatives themselves, not just their copy.

ask things like
  • Which hooks does this competitor repeat across creatives?
  • Cluster the ads by angle, offer, and audience.
  • Which creatives ran longest — and what do they share?
  • Compare EU reach by country against the stated targeting.
  • Draft three new variants in their best-performing pattern.

Install

~2 minutes, no store

  1. Download and unzip

    Grab adracle‑chrome.zip and unzip it somewhere permanent — for example ~/Extensions/adracle/.

    Chrome will load the extension from this folder, live. If you delete or move it later, the extension breaks — so don’t leave it in Downloads.

  2. Turn on Developer mode

    Open chrome://extensions and flip the Developer mode toggle in the top-right corner. This is the only way Chrome installs extensions from outside the Web Store — no registry hacks, no special builds.

  3. Load unpacked

    Click Load unpacked and select the folder you unzipped. The adracle card appears in your extensions list, and the “ad” tile lands in your toolbar (pin it via the puzzle-piece menu).

    Chrome may show a “developer mode extensions” notice on next launch — that’s expected; click it away.

  4. Accept the risk notice

    Click the adracle icon to open the side panel. It opens on Settings the first time — read the Terms-of-Service notice and tick the acknowledgement. Crawls stay disabled until you do.

Use

crawl → enrich → export

  1. Open the Ad Library and shape your query

    Go to facebook.com/ads/library and filter as usual — pick an advertiser page or a search term, set country, status, media type, sort. The URL in the address bar is the query; adracle crawls exactly what you see.

  2. Start a job from the side panel

    Open the side panel: the URL field already follows your active Ad Library tab, including filter changes. Optionally cap the crawl with an ad limit (e.g. 50), then hit Start. adracle scrolls the feed at a human pace and captures every ad that loads.

    Interruptions pause, they don’t fail — closing the tab mid-crawl keeps the data; hit Resume to continue from the checkpoint.

  3. Let the EU details phase run

    After the feed ends, adracle revisits each captured ad once (sequential, 2–4 s apart) to fetch its EU transparency block. One-time setup: open any single ad’s transparency panel by hand once, so adracle can learn the request shape — it tells you in the job log if this is still needed.

  4. Export

    When the job shows done, hit Export: one JSON file with every ad, snapshot, and EU block. It lands in AdLibrary/exports/ next to the media in AdLibrary/media/ — both inside whatever folder your browser saves downloads to. Browse what you’ve captured anytime in the Ads tab.

    Then hand the JSON to an AI — that’s what the format is for. See Built to be read by AI.