# creative intelligence tagging workbench - AdsMetri

Tag Meta ad creatives by hand and compare performance across 7 tag dimensions.

Type: project entry

Published: 2026-05-16

[Original page](<https://artlu.ai/project/creative-intelligence-tagging-workbench-adsmetri>)

[Browse this group](<https://artlu.ai/ai/groups/adsmetri.md>)

Parent project: AdsMetri

Recorded status: launched

Updated: 2026-05-20T19:37:39.331Z

Technologies: React, Tailwind CSS, shadcn/ui, tRPC, MySQL, Drizzle, Meta Graph API

## Details

**What is AdsMetri Creative Intelligence?**
A workbench for tagging Meta ad creatives by hand so you can compare performance across creative types, hooks, video structures, and visual styles.

**What it does**
Syncs ad-level creatives + daily performance snapshots from Meta. Shows them in a dense ops table with thumbnail, campaign / ad set, product, type/hook chips, status, spend, ROAS, CPA, CTR, purchases, tagged-state pill.

Click a row to open the tagging drawer — preview frame and external links on the left, 7 multi-select tag dimensions on the right, sticky Save + Next footer. External preview links cover the live Instagram and Facebook posts plus the Meta "share this ad" popover (two `fb.me/adspreview/*` URLs, one for Facebook-login viewers and one for managed-Meta accounts).

Keyboard flow: `↵` Save + Next, `S` Save, `K` Skip, `1-9` pick chip, `Esc` close.

Manage-tags page to rename, archive, merge per ad account.

**How it works**
Tags are stored as comma-separated slugs per dimension — multi-select round-trips through the existing single-string field, no schema break. External preview links come from Meta's `ad.previews` edge plus the creative's IG/FB permalinks. No inline video embed — Meta gates that permission, so the preview frame stays thumbnail-only with a small VIDEO badge.

**Built with**
React 19 · tRPC 11 · shadcn/Radix · Tailwind · Drizzle · MySQL · Meta Graph API v25.0

## Links

[Site or demo](<https://adsmetri.com/>)

An interactive demo is embedded on the original page.


Status and dates describe the published record, not a recent test of an external service. Missing details are not inferred. Public links do not imply an open-source licence or permission to reuse material.

[Directory](<https://artlu.ai/llms.txt>)
