The Help Centre: Search That Actually Finds Things
Fuzzy matching, synonyms, quoted phrases, whole-word highlights, shareable URLs, and keyboard navigation — the Help Centre nails every detail.
A help centre is the first thing every new user touches and the last thing most teams invest in. The Help Centre in this platform is not a generic FAQ list bolted onto the side; it is a small but carefully-designed search experience. The bet is simple: if a user can find what they need in five seconds, they will use the platform; if they cannot, they will guess and break things.
What is in the index
Every module ships with topic entries — quick how-tos, troubleshooting steps, and tips. Topics are tagged by module so the Help Centre can surface them next to the panel that needs them via the HelpButton component. The same index powers global search.
Fuzzy matching with synonyms
The search engine matches on prefixes, tolerates typos within a small Levenshtein budget, and expands query tokens against a domain dictionary so "rollback" finds "revert" and "backout". Synonyms are edited as data, not code; adding a new alias is a one-line change.
Quoted phrases
Multi-word terms in quotes are matched literally and weighted higher than the same words appearing separately. "verify all" is treated as one phrase; verify all is treated as two words that may match independently. This is the difference between a search box that looks smart and one that actually behaves the way users expect.
Highlights that respect words
Matches are highlighted at word boundaries. Searching for "ip" does not light up half of every paragraph; it lights up the actual occurrences. Phrase highlights are anchored on word starts and extended to word ends so the highlight aligns with what the user typed.
URL persistence
The active query lives in the URL. Refreshing preserves state; pasting the link into chat preserves state; bookmarking a search-by-tag is a real thing you can do. State that survives reload is state that earns trust.
Keyboard navigation
Enter jumps to the next match, Shift+Enter to the previous, arrows scroll the result list, and H toggles highlights for when you want to read without coloured spans everywhere. The shortcuts are non-obtrusive — they only fire when no editable field is focused — and they are the fastest way to triage results once you are familiar.
Help feedback
Every topic ends with a thumbs-up / thumbs-down. Negative feedback writes a row to help_feedback with the query that led to the topic, which gives the maintainers a real signal — "this topic shows up for these queries and is unhelpful" — instead of vague impressions.
Authoring topics
- Lead with the action. Users skim for verbs.
- Include the exact panel name and button label. Internal jargon costs trust.
- Add at least one synonym for any word a new user might use instead of the canonical term.
What to try next
Open Help, type a phrase you would naturally use during an incident — not the term the platform prefers — and see whether the right topic comes up. If it does not, the synonym dictionary needs another line. That is a one-PR fix, and it makes the next user's experience better.
Was this helpful?
Generate scripts for the task covered in this post
Describe what you need and NAPT writes production-ready automation in Netmiko, NAPALM, or Nornir.
Open Script Generator →Related posts
From Noise to Narrative: How the RCA Engine Builds a Causal Story From Your Findings
The Root Cause Analysis engine reads every finding your scripts produce and stitches them into a single causal chain — root, contributing, symptom.
Read →The Streaming Console: A Live Terminal That Remembers Everything
Run any diagnostic script against any device, watch output stream in real time, and end up with structured findings the platform can reason about.
Read →Intent: Saying What You Mean and Proving the Network Agrees
Write a one-line statement of how the network should behave. The Intent engine compiles it into checks and tells you where reality diverges.
Read →