Import Public Documents From A URL

Add useful public context to a project without copying it by hand. Snipara shows what it found before writing anything, then securely refetches the source when you explicitly import it.

How it works

1. Open a project's Context tab and choose Import URL.
2. Paste a public HTTPS URL and preview the source.
3. Review the title, host, detected format, extracted size, token estimate, sections, parser, and suggested project path. Nothing is saved yet.
4. Adjust the relative Markdown or text path if needed, then import.
5. Snipara securely refetches, deduplicates, saves, and reports whether indexing is queued, pending, or unnecessary.

Supported sources

Direct document links use Snipara's existing deterministic parser lane. A page or file still needs to be publicly reachable and produce readable text within the importer limits.

Safety and privacy

• Only public HTTPS on port 443 is accepted.
• Every DNS result and redirect destination is checked before access.
• Fetch, redirect, parser, archive, and extracted-content limits are enforced.
• Preview is side-effect-free; commit never trusts browser-returned content.
• Stored provenance excludes URL queries and fragments.
• Errors do not expose private addresses, credentials, bodies, or raw exceptions.

Duplicates and indexing

Importing the same source again refreshes its existing path. If identical content already exists elsewhere, Snipara returns that duplicate instead of silently creating another document. A saved document and a fully indexed document are different states, so the result reports indexing honestly.