Social Metadata Generator

Social Media Metadata Generator

Give it a video topic or a full script and it returns a complete publishing metadata pack: several title options, an optimized description/caption, hashtags, SEO keyword tags, and a pinned-comment idea. Everything is tuned to one platform at a time (YouTube long-form, YouTube Shorts, TikTok, Instagram Reels, or X/Twitter) so the character limits and hashtag counts match where you're actually posting. It's meant for people running a content pipeline who want to stop hand-writing titles and tags for every upload.

How it works

Each item goes through a single chat-completion call with a platform-specific prompt that enforces that platform's title length, description limit, and hashtag count. The response comes back as structured JSON, then gets cleaned up (hashtags normalized to #nospaces, tags stripped of #).

Input

You need either text or sources. Everything else has a sensible default. An OpenAI key is required for real generation because it runs on an AI model. If you run it with no key, the actor returns one clearly-labeled sample record (_sample: true) so you can preview the output shape — add your key for real, AI-written metadata.

Output

One dataset record per input item. Each record has platform, titles (array), description, hashtags (array, #-prefixed, no spaces), tags (array of bare SEO keywords), and pinnedComment. Records also carry ok and the truncated source; if a single item fails the run keeps going and that record gets ok: false with an error.

Example

{
  "text": "A 60-second video explaining why cats knock things off tables, with funny examples.",
  "platform": "youtube shorts",
  "titleCount": 5,
  "keywords": "cat behavior, funny cats, pet tips",
  "language": "English",
  "openaiApiKey": "sk-..."
}

Pricing

$0.015 per metadata pack — $15.00 per 1,000 — plus a $0.0000125 run-start fee each time a run begins. Flat rate: no volume tiers, no plan gates, no subscription.

One pack is the whole set for one item: every title variation, the description, the hashtags, the SEO tags and the pinned comment. Raising titleCount from 5 to 15 does not raise the price. Because sources takes a batch, a run that generates packs for 40 uploads costs $0.60 plus a single fraction-of-a-cent start fee. You also pay your own OpenAI usage on your own key, which is billed by OpenAI and not marked up here.

Notes

A small titleCount (5) is a fine default. Pushing it to 15 spends more tokens for variations that start to repeat themselves. Pick the right platform before running, since switching it changes the limits the model writes to, not just a label.