PodHood is in private beta — request access.
PodHoodDocs

Syncing & indexing

The mental model — source of truth, what syncs, what gets indexed, and what it costs.

Two separate machines run under every Channel. Sync keeps your catalog mirrored; indexing makes episodes intelligent. Knowing which is which explains almost everything about how PodHood behaves.

The source of truth

Each Channel has exactly one source — a YouTube channel or an RSS feed. PodHood mirrors it and never edits it: you keep curating where you already publish, and the source always wins. If something changes at the source (a title, artwork, a deleted episode), PodHood follows.

Sync — the mirror

Sync brings your catalog in: episode rows, titles, artwork, collections. It's cheap, automatic, and continuous — new episodes are picked up within about an hour of publishing, with no action from you.

A few behaviors worth knowing:

  • Deletions archive, never destroy. An episode that disappears from your source is archived on PodHood, not hard-deleted.
  • Address drift is safe. Episode identity is keyed on stable ids (a YouTube video id, an RSS <guid>), not on file URLs — migrating podcast hosts or rotating enclosure URLs won't orphan your Library or re-charge transcription.
  • One ghost download per episode. Transcription fetches your audio file once, so hosting analytics will record roughly one extra download per episode.

Indexing — the intelligence

Indexing runs one episode through the AI pipeline: transcription (word-level timing), then the Brief — summary, chapters, key moments, speakers with roles, entities, topics. The first indexing is what publishes the episode to your public Library; un-indexed episodes stay out of it.

The transcript is treated as stable and costly: produced once, reused forever. A reindex (refreshing a Brief under newer models) never re-transcribes.

Backfill — your back catalog

Importing mirrors your whole catalog, but only your newest episodes are indexed automatically. Making the historical archive searchable and citable is a backfill — bulk indexing of everything older. Same pipeline, same rate; it differs from ongoing indexing only in timing.

What it costs

Indexing is the metered thing (sync is free). Plans grant credits — one credit is one hour of processed audio or video — and backfill and new-episode indexing draw from the same pool. Current early-access plans and rates are on the pricing page.

Questions teams actually ask

Does PodHood replace my podcast host or website? No. It's a mirror plus an index — your feed, host, and site stay exactly as they are. The Library is an additional surface (which you can put on your own domain).

What happens when I fix a title or thumbnail at the source? The next sync picks it up. There's nothing to edit on PodHood — deliberately.

Can I choose what gets published? Publishing follows indexing. Index an episode and it's live in the Library; an archived-at-source episode leaves it.

What language does my Library speak? The Channel has a Podcast language setting (Settings → General). It's the default for transcript translation and sets the public Library's display language.

Was this page helpful?

On this page