Drive videos posted to TikTok and Instagram with AI captions

Drive videos posted to TikTok and Instagram with AI captions

An n8n workflow that waits for a new video in a Google Drive folder and transcribes its audio with OpenAI. A second OpenAI call writes a post description from the transcript, and the video goes to TikTok and Instagram with that text through an upload service, Upload-Post.

What is inside

• workflow.json: the workflow as n8n exports it, thirteen nodes: twelve that do the work and one sticky note. It holds no credential.
• README.md: these notes.

What we changed from the original copy

• The credential links are gone. Select your own for Google Drive, OpenAI, Telegram and the two Header Auth steps.
• The description step used gpt-4o. OpenAI's deprecations page, read on 12 November 2026, lists the dated snapshots of that model as shut down on 14 January 2027 with gpt-6-luna as the replacement. The step has no tools, so this copy names gpt-6-luna.
• The three example descriptions in the prompt promoted a product by name. They are replaced with three neutral ones. Swap in your own best posts.
• The private ids are placeholders: <YOUR_FOLDER_ID>, <YOUR_TELEGRAM_CHAT_ID> and <YOUR_UPLOAD_POST_USER>. The notes of the two upload nodes show the header to use, with <YOUR_UPLOAD_POST_API_KEY>.
• The download node's file id had a blank mode. It is set to By ID.
• The If node read $json.trigger.error.message. n8n's Error Trigger page says trigger holds data only when the error is in the workflow's trigger node. So an error in a later node made the check itself fail, and no alert went out. It now reads that message when there is one, and an empty text otherwise.
• The big note is rewritten as a setup list.
• Nothing else.

Before you start

• An n8n instance, on n8n Cloud or your own server, that may write files. Two nodes save the video to disk. If you limit file access with N8N_RESTRICT_FILE_ACCESS_TO, the folder must be on that list.
• A Google account for Drive, and an OpenAI API key.
• An Upload-Post account with TikTok and Instagram connected, an API key and the name of your profile there.
• Optional: a Telegram bot token and your chat id, for error alerts.

Set it up

1. In n8n, create a workflow, open its menu and choose Import from File. Pick workflow.json.
2. Add a Google Drive credential and select it in Google Drive Trigger and Google Drive. Replace <YOUR_FOLDER_ID> in the trigger.
3. Add an OpenAI credential and select it in Get Audio from Video and Generate Description for Videos in Tiktok and Instagram.
4. Add a Header Auth credential with the name Authorization and the value Apikey <YOUR_UPLOAD_POST_API_KEY>, with your key in place of the placeholder. Select it in both upload nodes, and replace <YOUR_UPLOAD_POST_USER> in each with your profile name.
5. Edit the examples in the prompt of Generate Description for Videos in Tiktok and Instagram.
6. Optional: add a Telegram credential to Telegram and replace <YOUR_TELEGRAM_CHAT_ID>. A workflow that holds an Error Trigger uses itself as its error workflow by default, and the trigger runs only for automatic runs.
7. Activate the workflow, then add a short test video to the folder.

How a run goes

1. Google Drive Trigger checks the folder every minute for a created file.
2. Google Drive downloads the file into the binary field data.
3. Read video from Google Drive is a Write Binary File node. It saves the video under its original name with spaces turned into underscores. The path has no folder in it.
4. Get Audio from Video sends the data field to OpenAI for a transcript.
5. Generate Description for Videos in Tiktok and Instagram writes the description from the transcript. The prompt asks for the description alone.
6. Two Read Binary File nodes, both named Read Video from Google Drive apart from a 2, read the saved file back into the field datavideo, one for each platform.
7. The two upload nodes post a form to Upload-Post: title with the description, minus double quotes, platform[] set to tiktok or instagram, the video and the user.
8. Separately, Error Trigger feeds If, which sends the Telegram text 🔔 ERROR SUBIENDO VIDEOS unless the error is the DNS message about a server being offline.

Change the look

• The tone of the description: the prompt and its examples.
• The platforms: Upload-Post lists more, and the platform[] value of each upload node picks one.
• The Telegram text: Telegram.
• How often Drive is checked: Google Drive Trigger.

Good to know

• The workflow publishes to real accounts as soon as a video appears. Try it with test accounts first.
• Nothing deletes the saved copies, so they pile up on the n8n server.
• whisper-1 is named inside n8n's audio step. OpenAI's deprecations page lists it for shutdown on 26 February 2027 with gpt-live-transcribe or gpt-transcribe as replacements. This copy leaves it as the file has it. Swap that node for an HTTP Request to /v1/audio/transcriptions before that date.
• OpenAI's speech-to-text guide gives a limit of 25 MB for a file, so a larger video fails at the transcript.
• n8n's source marks Write Binary File and Read Binary File as hidden, so the node panel does not offer them for new workflows, but n8n's package still lists them.
• The video and its description go to Upload-Post, which posts them for you under its own terms.
• This copy was not run against Google Drive, OpenAI, Upload-Post or Telegram.
• The JSON holds no credential.