Skip to content

Self-Hosted Getting Started

This is the Self-Hosted path through InKCre: you will operate your own instance. For the application's What, Why, and How, begin with Getting Started.

You will use the Web app for interactive setup. A trusted Agent or operator can handle the remaining terminal-only operations, such as installing Core collectors and scheduling indexing. No InKCre development setup or AI model is needed for the first collection-and-search journey. Each guide below is self-contained and can also be used with an existing compatible instance.

1. Deploy an instance

Choose one deployment guide:

Neon and the compute providers are convenience options, not requirements. At the end, retain a ready Core URL, a PostgREST URL, your private JWT secret, and the Core Peer ID.

2. Complete the first collection-and-search loop

Follow these shared user guides in order. If you already have an instance, start here.

  1. Connect to your instance, using the Web app or the CLI for your Agent.
  2. Collect your first source; a public RSS or Atom feed is a simple starting point.
  3. Find what you saved: maintain the lexical index, search, and inspect a result.

Checkpoint: you can retrieve a real item from your source. Collection and indexing are separate; neither requires you to set up AI organization first.

3. Make it useful day to day

These paths make information available on demand. They do not configure automatic daily digests or Telegram/email push notifications. Hosting that sleeps cannot guarantee continuous collection.

If a step fails, use Troubleshooting. For backups, upgrades, and ongoing hosting responsibilities, return to Advanced Self-Hosting.