What to expect
Indelible aims to be honest about what it does today. Here’s the current scope so there are no surprises.
Capture
Section titled “Capture”- Articles, newsletters, PDFs, and EPUBs are archived in full. The text and images are fetched and stored at save time, so your copy survives even if the original disappears.
- Videos are saved with their transcript and metadata, not the media file. For a YouTube or other video page, Indelible captures the transcript (where one is available) and details like title and channel; it does not download or host the video itself.
- Podcasts are not a first-class content type at launch. Podcast RSS entries can still arrive through the ordinary feed reader, where their show notes are treated like articles. Indelible does not persist or play enclosure audio.
- Podcast transcripts from Spotify and Apple Podcasts are not supported. Indelible does not extract transcripts from either service at this time.
Hosting
Section titled “Hosting”- Indelible Cloud is coming soon. Today, Indelible is self-hosted; run it with Docker on your own hardware. See Install with Docker.
- The web interface is served separately from the backend images. It’s a static single-page app you host alongside your API. A dedicated web-hosting guide is coming.
Operational notes
Section titled “Operational notes”- Rate limiting is per-instance. Single-node deployments are fully supported; coordinated rate limiting across multiple API replicas is not wired up yet.
- Inbound email needs a provider. Saving by email requires configuring an email provider (Resend) and the matching DNS, not just an address.
- Mila and semantic search need an AI provider. Full-text search works out of the box; semantic search and the Mila assistant require configuring an embeddings/chat provider.
- Reading EPUBs and PDFs on mobile is coming next. The mobile apps don’t open EPUB or PDF files yet; that lands in the next iteration.