The Knowledge Base is a list of URLs that TxID Support indexes and searches when answering questions. When a user asks something the AI can't answer from on-chain data alone, such as how your governance works, what your tokenomics are, or how to bridge, it searches your indexed pages to find the answer.
How indexing works
When you add a URL, TxID Support crawls that page, extracts the main text content, and splits it into searchable chunks. When a user asks a question, the AI retrieves the most relevant chunks and uses them to compose a response. It can also cite the source page so users can read more.
What to index
Index any public page that answers questions your users are likely to ask:
- Protocol documentation: how staking works, how to bridge, how to claim rewards
- Tokenomics pages: supply, distribution breakdown, vesting schedule overview
- FAQ pages: existing common questions and their answers
- Governance documentation: how proposals work, how to vote, quorum requirements
- Security information: audit report summaries, multisig setup, emergency procedures
- Roadmap pages: what's live, what's coming, key upcoming dates
Quality beats quantity
Ten well-written, detailed documentation pages outperform fifty thin or duplicated pages. The AI retrieves chunks semantically. If your docs repeat the same information across many pages, it may retrieve lower-quality matches.
Adding a URL
Go to Docs & KB
Click Docs & KB in the dashboard sidebar.
Paste a URL
Enter the full URL of the page you want indexed (e.g. https://docs.yourprotocol.io/staking).
Click Index
TxID Support crawls the page and adds the chunks to your knowledge base. Most pages index in under 30 seconds.
Repeat for each page
Add all the key pages from your documentation. The indexed chunk count on your dashboard overview shows the running total.
Re-indexing
Indexed content is a snapshot taken at crawl time. If your documentation changes, re-index the affected URLs to keep the AI's knowledge current. The old chunks are replaced with the new content.
Public pages only
The crawler can only access publicly available pages. Content behind a login, paywall, or IP restriction cannot be indexed.
Tips for better coverage
- Index specific subpages rather than just your homepage (the crawler doesn't follow links automatically)
- Include pages that explain your core concepts in plain language, not just technical reference
- A single long page is fine. It will be split into searchable chunks
- Avoid indexing pages that are mostly navigation, headers, or boilerplate: they add noise without useful content
- If your docs are updated frequently, schedule a regular re-index (monthly is usually sufficient)