See llms.txt for all machine-readable content.

Back to Templates

Create and review multilingual avatar video campaigns with HeyGen and OpenAI

Created by

Created by: Andy || banandyy
Andy

Last update

Last update 3 days ago

Categories

Quick overview

Create a vertical HeyGen avatar master from an approved script, review target-language subtitles before final translation, and archive verified videos and captions privately in Google Drive. Separate language approvals, original provider IDs, and OpenAI social-copy drafts help teams prepare multilingual campaigns for human review.

How it works

  1. Reads approved campaign briefs from Google Sheets, validates authorized avatar/voice assertions and target languages, and registers an immutable snapshot in a native Data Table. Duplicate IDs and changed registered inputs stop processing.
  2. Persists an exclusive exact-payload intent before creating a clean vertical master through HeyGen V3. It polls the original ID finitely, inspects downloaded MP4 container/audio metadata and archives the actual bytes in a private Drive folder.
  3. An allowlisted signed-in n8n user reviews the exact saved master version and hash. Only master approval unlocks an independent V3 proofread session for each configured target language.
  4. Downloads the actual target-language SRT for review. Corrections require reviewed SRT text and a matching immutable HTTPS file.
  5. It compares downloaded cues, uploads the correction through the documented V3 URL contract, and compares the provider readback before language approval.
  6. Creates final translations only after exact per-language approval, retaining separate proofread and translation IDs. Downloads each final MP4, SRT and VTT, checks actual media structure, target timings and approved transcript consistency, and archives verified bytes privately.
  7. Uses OpenAI to draft an accompanying title and social caption from the approved target transcript. A final exact-version review checks the actual translated files and copy. Google Sheets mirrors durable status; this workflow does not publish or schedule videos.

Setup

  1. Import the setup, worker and authenticated review companion from the purchase package. Run the credential-free local synthetic demo first. Keep production workflows inactive and generationEnabled=false while configuring.
  2. Connect Google Sheets and Drive OAuth2 in setup, then create fresh Campaigns, Data Table and private Drive resources. Copy all returned IDs into the worker and review policy nodes. Use matching policy values and the documented exact headers.
  3. Configure actual allowlisted n8n reviewer emails, permitted private-folder user emails, exact trusted HTTPS asset hosts, supported target-language names and an OpenAI model. Keep n8n User authentication enabled on the review form and the archive folder private.
  4. Connect HeyGen X-Api-Key Header Auth, OpenAI, Google Sheets and Drive credentials to their respective nodes. Supply existing authorized V3 photo-avatar and voice IDs; this package does not create new avatars or clone voices. Confirm actual presenter, voice and content rights.
  5. Add a new unique campaign ID and approved script. Set rights_confirmed and approved_for_master TRUE only after review, then deliberately enable generation. Run the worker manually to register and perform successive operations. Preview the exact saved version/hash through the review form and separately approve the master, target SRTs and final files.

If correcting subtitles, provide both complete corrected SRT text and a stable provider-accessible HTTPS URL containing identical cues. Review provider costs and access before enabling paid work. Activate the minute schedule only after deployment-specific tests.

Requirements

  • n8n with native Data Tables, Forms with n8n User authentication, HTTP Request, Code, Google Sheets, Google Drive and Compression nodes; Google OAuth2 access to Sheets and private My Drive; HeyGen API access for V3 avatar videos and speed translation/proofread sessions, an existing authorized photo-avatar ID and voice ID; an OpenAI API credential; exact trusted HTTPS asset hosts and stable correction-file hosting when editing SRTs; actual presenter/content rights and human target-language reviewers. Up to two target languages per campaign. The template price excludes hosting, provider credits/tokens and storage. No community nodes or runtime command-line tools are used.

Customization

  • Change supported target languages, approved source script, authorized avatar/voice, brand terms, OpenAI copy model, exact trusted hosts, reviewer/folder-user allowlists and documented finite limits in the matching policies. Use a new campaign ID for changed registered creative inputs. Keep separate per-language SRT approval, final-file review, private permissions and durable request hashes. Extend with a separately reviewed publishing workflow if needed; this package does not post or schedule social content.

Additional info

The package includes worker, authenticated review, setup and credential-free synthetic demo JSONs, exact queue/schema examples and test notes. Exports are inactive with generation disabled and no bound credentials. Local synthetic/media/state tests and isolated native mocks do not prove real HeyGen rendering, avatar consent permissions, translation quality, lip sync, live Google access or n8n Cloud behavior. Synthetic examples contain colored frames and tones, not real avatar speech. Watch and check each actual language output before approving. Unknown mutation outcomes stop for original-ID reconciliation rather than blind retries; HeyGen's 24-hour idempotency replay is guarded by exact persisted payload hashes. Missing caption URLs stop the affected language, without reusing source timings. Third-party services can bill separately. No publication is performed.