CONTENT CREATION

Agent-run cut-out QA with auto-reshoot on rejects

An agent reviews freshly cut-out product images for edge artifacts and color cast, re-runs failures through an alternate model.

CategoryContent Creation
Enginepaperclip
Difficultyadvanced
Triggerevent
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew cleaned PNG lands in S3AWS S3
  • LogicAgent scores mask quality pass or fail
  • ActionRe-run failures through alternate Replicate modelReplicateReplicate
  • OutputWrite passing assets to Notion publish-ready catalogNotionNotion
  • OutputFlag persistent failures to Slack for reshootSlack

What it does

Adds a quality gate to your cut-out pipeline. An agent inspects each transparent PNG, decides whether the masking is clean, and automatically retries weak results with a different Replicate model before anything reaches your live catalog.

When to use it

Use it when cut-out quality varies by product type (fuzzy, reflective, or thin items break naive masking) and shipping a bad edge to the storefront is costly. It replaces a human spot-checking every export.

How it works

  1. 1A new cleaned PNG arriving in cloud storage triggers the agent.
  2. 2The agent evaluates the image for halo edges, leftover background, and color cast, scoring pass or fail.
  3. 3On a fail, it re-runs the original through an alternate Replicate model and re-evaluates, up to a retry limit.
  4. 4Passing images are written to the Notion publish-ready catalog database with their QA score and notes.
  5. 5Persistent failures are flagged to a Slack channel for manual reshoot, so nothing silently disappears.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect AWS S3Buckets, objects, signed URLs.
  2. 2
    Connect ReplicateImage, video, and model inference.
  3. 3
    Connect NotionPages, databases, comments.
  4. 4
    Connect SlackChannels, DMs, threads, mentions.
  5. 5
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  6. 6
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  7. 7
    Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.

Run this workflow in your colony.

14-day trial. No DevOps. No Sales call. Provisioned in under a minute.