CONTENT CREATION

Upscale and caption low-res screenshots into annotated set

Takes low-resolution or compressed screenshots, upscales them with a Replicate model, generates a descriptive caption per image with OpenAI, burns the captions into Figma callouts.

CategoryContent Creation
Enginesim
Difficultyintermediate
Triggerevent
Steps5
Setup~15 min

How it runs

The automated pipeline, trigger to output.

  • TriggerNew low-res image in R2 ingest prefixCloudflareCloudflare R2
  • ActionUpscale image with Replicate super-resolution modelReplicateReplicate
  • ActionGenerate instructional caption from imageOpenAI
  • ActionCompose upscaled image and caption into Figma frame and exportFigmaFigma
  • OutputWrite polished annotated image to R2 published prefixCloudflareCloudflare R2

What it does

Rescues low-quality screenshots and turns them into a clean annotated tutorial set. It runs each image through a Replicate super-resolution model to sharpen it, asks OpenAI to write a short instructional caption describing what the screen shows, places the upscaled image plus caption into a Figma callout layout, and saves the result to R2.

When to use it

Use it when the only screenshots you have are blurry, downscaled, or pulled from old docs, and you need them to look crisp and carry clear captions before republishing.

How it works

  1. 1A new image lands in the watched R2 ingest prefix and fires the trigger.
  2. 2Replicate upscales the image with a super-resolution model and returns a high-res render.
  3. 3OpenAI generates a concise instructional caption from the upscaled image.
  4. 4The flow composes the upscaled image and caption into a Figma callout frame and exports a PNG.
  5. 5The finished annotated image is written back to R2 under the published prefix.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect Cloudflare R2Object storage, S3-compatible.
  2. 2
    Connect ReplicateImage, video, and model inference.
  3. 3
    Connect OpenAIModels, embeddings, files.
  4. 4
    Connect FigmaFiles, frames, comments, assets.
  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.