CONTENT CREATION

Localize annotated tutorial image sets into multiple languages

Takes an annotated English screenshot set, translates every callout label into your target locales, re-renders the labels onto each image in Figma.

CategoryContent Creation
Enginesim
Difficultyadvanced
Triggermanual
Steps5
Setup~25 min

How it runs

The automated pipeline, trigger to output.

  • TriggerManual run with Figma file key and target locales
  • ActionExtract callout text layers from Figma masterFigmaFigma
  • ActionTranslate label strings per localeOpenAI
  • ActionRewrite text layers and export PNG set per localeFigmaFigma
  • OutputPublish locale-prefixed image sets to R2CloudflareCloudflare R2

What it does

Produces localized copies of an annotated screenshot tutorial. Starting from a master Figma annotation file whose callout labels are real text layers, it translates each label into a list of target languages, swaps the text per locale, re-exports the images, and ships one complete image set per language to R2 so your localized help center pulls them directly.

When to use it

Use it when a tutorial ships in English first and you need de, fr, es, and ja versions of the same annotated screenshots without redrawing arrows for every locale. The geometry stays fixed; only label text changes.

How it works

  1. 1You trigger the run with a Figma file key and the list of target locales.
  2. 2The flow pulls every callout text layer and its string content from the Figma file.
  3. 3OpenAI translates the label strings as a batch per locale, preserving UI terminology.
  4. 4For each locale the flow rewrites the text layers in a duplicated Figma page and exports all frames as PNGs.
  5. 5Each language set is uploaded to R2 under a locale-prefixed path, ready for the docs CDN.

Set it up

What you configure once, before turning it on.

  1. 1
    Connect FigmaFiles, frames, comments, assets.
  2. 2
    Connect OpenAIModels, embeddings, files.
  3. 3
    Connect Cloudflare R2Object storage, S3-compatible.
  4. 4
    Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
  5. 5
    Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
  6. 6
    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.