Quick answers to the questions teams ask most about StretchSearch. For deeper walkthroughs, follow the links in each answer to the relevant guide or feature article.
Basics
What is StretchSearch, in one sentence? A governance-first platform for crawling, indexing, and ranking site content, and for grounding search and AI in real, citable evidence.
Is it like Algolia or a hosted search box? It solves the same "make content searchable" problem, but inverts the priorities: authorization, robots policy, risk, rate limits, and approval come before any crawl. It's built for teams who need to defend what they crawled.
Do I need admin rights? No for everyday work — projects, targets, crawl jobs, readiness, searches, and reviewing findings/rankings/evidence are all workspace actions. Generating findings or ranking previews and toggling the kill switch are operator-only, governed actions.
Crawling and governance
Does creating a target or crawl job start a crawl? No. It creates a governed record. Crawler execution is disabled by default (crawler_enabled: false) and is enabled centrally only after authorization, readiness, and approval.
Why does every crawl response say "no crawler execution was started"? Because that's true and intentional. StretchSearch separates planning and governance from execution. See The Authorization, Governance & Safety Model.
What controls whether a job can run? Three things must all be satisfied: the readiness checklist (required items passed), governance approval (approved), and centrally enabled execution — and the master kill switch must not be engaged.
What are the safety limits? Per job: max_depth 0–10, max_pages 1–5,000, rate_limit_ms 250–60,000, plus respect_robots and a crawl scope from seed_urls_only to approved_deep_crawl.
What is the kill switch? A master control that halts all crawler execution instantly. Anyone can read its status; toggling it is a governed, audited operator action. See Crawl Audit Monitoring and the Kill Switch.
Search
What does federated search cover? Projects, indexed pages, crawl targets, crawl jobs, SEO findings, and ranking previews — across StretchSearch, StretchProjects, the crawler, and StretchSEO — in one ranked, faceted list. See Running Federated Search.
How are results ranked? By lexical_freshness_authority_v1, blending lexical match, a base score per type, freshness (decaying over ~90 days), and authority, with an optional semantic re-rank on top.
Can I save searches? Yes — a query is required, and you can name and reuse saved searches. Recent searches are remembered automatically. Both are scoped to your organization.
SEO findings
Where do SEO findings come from? A deterministic engine analyzes indexed pages for issues like missing titles (high), non-200 pages (high), missing meta descriptions and H1s (medium), thin content (medium), and duplicate titles/metas. Same pages → same findings. See Reviewing SEO Intelligence Findings.
What's the SEO health score? A 0–1 roll-up that weights findings by severity against the number of pages. Higher is better.
I fixed an issue but it still shows open — why? Findings reflect the last generation. Re-generating supersedes the old open findings with a fresh set.
Candidates, ranking, and evidence
What is candidate indexing? The extraction of keywords, phrases, entities, page chunks, and chunk references from indexed pages — the raw material for ranking and retrieval. See Candidate Indexing Explained.
How does ranking work? The deterministic StretchRank engine scores candidates by type weight (entities > phrases > keywords > chunk refs), frequency, references, and evidence, minus a quality penalty — with every score component exposed. Ranking previews are currently a governed preview surface. See Understanding Evidence Ranking Previews.
What are evidence references? Citable records (from pages, findings, AI runs, or notes) with a lifecycle status (active, stale, superseded, revoked) and a citable flag. They ground findings, rankings, and AI answers. See Evidence References and Citations.
How does StretchSearch ground AI? Through retrieval context (relevant evidence chunks for a query) and the evidence package that bundles retrieval plus ranking signals for AI surfaces like StretchGPT. See Retrieval Context and the AI Evidence Bridge.
Suite fit
How does StretchSearch relate to StretchBAS? StretchSearch is infrastructure — the crawl/index/evidence layer. StretchBAS is the brand-audit and reporting product that consumes StretchSearch's evidence. StretchSEM/SEO consume its findings and candidates, and AI surfaces consume its evidence bridge.
Is my data isolated from other organizations? Yes. Projects, pages, findings, searches, and evidence are scoped to your organization.
Still stuck?
If something isn't behaving as expected, start with Troubleshooting Common StretchSearch Errors, which covers validation messages, not-found responses, governance blocks, and empty results. For anything governed (generation or the kill switch), your platform operator holds the controls.
Was this helpful?
Help us improve this article
Use these controls to share whether this answer solved the issue. Feedback helps prioritize updates to StretchSuite Support.