H Habla Agente docs
/

Your product

Knowledge

Knowledge is what your Habla agent uses to answer visitors. You manage it from an agent’s Knowledge screen (or from onboarding’s Index your website step).

Source vs page

  • Source — a website or other import you add and manage in Knowledge (for example one website, one pasted FAQ, one upload).
  • Page — an individual discovered or indexed URL that belongs to that source.

A source is the thing you create and reindex. Pages are the URLs that source discovers and indexes under it.

One website → one source

Website onboarding creates one website/sitemap source for your site. That single source can fan out into many indexed pages. Prefer one website source for a whole site — not one source per URL — unless you intentionally want a single page.

Sitemap vs no sitemap

  • Usable sitemap found — Habla can discover multiple pages from that sitemap and index them under the one website source.
  • No usable sitemap — Habla falls back to indexing the homepage only. You can add individual pages or a sitemap later from Knowledge.

Pages indexed vs plan allowance

  • Pages indexed — how many pages have been successfully indexed so far for the source (or across progress you see in the UI).
  • Plan page allowance — the maximum pages available to your workspace according to its Habla plan. Indexing will not keep expanding past that allowance.

Plan names, prices, and current allowances are shown in Plans & Billing and in /app/billing — this page does not list quotas.

Per-source infrastructure cap

Separately from your plan allowance, Habla enforces an infrastructure ceiling per source: a single source can only process up to a fixed number of pages, even if your plan allowance is higher. If indexing stops earlier than your plan suggests, this cap (or a missing sitemap) is a common reason. How crawling works under the hood: Crawler chain.

Indexing status

After you add or reindex a source, watch its status in Knowledge:

  • While work is in progress, the source is still being crawled or indexed — do not treat answers as fully grounded yet.
  • When the source shows as indexed, Habla has finished the current indexing pass for that source (for a website source, that means the pages it was able to process in this run — subject to sitemap discovery, plan allowance, and the per-source infrastructure cap).
  • If something fails, fix the URL or content and use Reindex (or edit pasted text and save).

Accepting a URL is not the same as having searchable indexed content. Rely on the Knowledge status and page counts before you depend on answers in production.

Reindex

Use Reindex when your website or imported content changed and you want Habla to refresh what the agent can cite. For pasted-text sources, prefer Edit when you need to change the text itself; Reindex refreshes from the saved body. After you start a reindex, wait until status settles again before expecting updated answers.

Other ways to add knowledge

Besides a website source, you can add single URLs, pasted text, file uploads, connected apps (such as Notion or Google), WordPress sync, and auto-indexed pages from visitor traffic. Choose the type that matches how your content lives.

Engine details
Crawl mechanics, retrieval, and data shape are documented in Crawler chain, RAG pipeline, and Data model — not restated here.

Related