n8n Workflow: Narrate Google Slides with Gemini TTS & Shotstack

Narrate Google Slides automatically—with Gemini TTS, smart re-voicing, and Shotstack-rendered captions

This n8n workflow turns your Google Slides speaker notes into a narrated video. It re-voices only the slides that changed, uses Gemini to draft the narration and suggest pronunciations, and then renders captioned MP4 videos via Shotstack—uploading the final results to Google Drive.

What this workflow does

• Starts from an n8n Form submission (with a Google Slides URL and options) or a scheduled weekday 8:00 run.
• Loads the presentation and fingerprints each non-skipped slide (including speaker notes and voice settings).
• Detects changes by comparing against stored slide records in an n8n Data Table, so unchanged slides are reused.
• Rebuilds narration for each changed slide using a Google Gemini agent that rewrites notes into a narration script and flags issues.
• Handles pronunciations by looking up reading dictionary entries in an n8n Data Table tool, generating proposed pronunciations when needed.
• Generates audio with Gemini TTS, uploads slide image + audio to Shotstack, and upserts updated scripts/captions/durations/asset URLs back into the slide Data Table.
• Routes review to Slack for approval/rejection.
• Renders & publishes an edit timeline in Shotstack (reusing unchanged assets), renders the final captioned MP4, uploads to Google Drive, and logs the render back to the Data Table.

Use cases

• Update training or product deck videos whenever speaker notes change—without re-rendering everything.
• Maintain consistent narration pronunciation using a shared dictionary and Gemini TTS voices.
• Give reviewers a Slack-based approval step before costly Shotstack renders.

Technical details

• Built with n8n nodes such as if, set, code, no op, wait, and slack.
• Integrates Google Slides, Google Gemini (agent + TTS), Slack, Shotstack, and Google Drive, using n8n Data Tables for slide change detection and pronunciation/script storage.