Human request
Original experience
SigilEdge retrieves and returns the origin response without applying AI enrichment to ordinary visitor traffic.
visitor→edge→origin HTML
Request lifecycle
From your origin to approved, bot-facing content
SigilEdge adds a controlled enrichment path for recognized AI crawlers. People continue to receive your original site.
pass through
Recognized crawler
Approved enrichment
When approved content is cached, SigilEdge merges it into the crawler response. On a miss, generation is queued without inventing a synchronous result.
crawler→edge→cache/review→enriched HTML
The six-stage lifecycle
Generation, approval, and delivery are separate stages so a slow model call or low-confidence output never has to block a live request.
Route
After domain ownership and origin configuration are complete, requests pass through the SigilEdge edge worker.
Classify
Known AI crawler user agents enter the enrichment path. Unknown or missing user agents receive the normal human path.
Fetch
SigilEdge retrieves the origin response. The origin must return meaningful HTML for enrichment to work well.
Generate
On a cache miss, the original response can be served while structured data, a summary, and FAQs are generated asynchronously.
Review
Confidence, claim-risk, and entailment rules determine whether content can publish or requires a human decision.
Deliver
Approved enrichment is cached and merged into later recognized crawler responses at the edge.
What gets added
Actual enrichment—not fictional citation tags
The output is grounded in standard web formats and the source page. SigilEdge does not add proprietary citation meta tags or guarantee how a third-party answer engine will use the content.
Schema.org JSON-LD
WebPage, Article, FAQPage, Product, Offer, AggregateRating, Breadcrumb, and Speakable where applicable.
Grounded summaries
Concise page summaries with source evidence, confidence, and an entailment check before automatic approval.
Grounded FAQs
Question-and-answer candidates generated from page content, with volatile claims routed to review.
Cleaner bot-facing HTML
Enrichment is added to the crawler response while the human-facing origin experience stays intact.
Crawler recognition
A maintained allowlist of real user agents
Unknown or missing user agents use the normal human path. Google-Extended is a robots control token, not a crawler user agent, so Gemini is not presented as a detected crawler.
OpenAI
GPTBot · ChatGPT-User · OAI-SearchBot
Anthropic
ClaudeBot · Claude-User · Claude-SearchBot
Perplexity
PerplexityBot · Perplexity-User
Other recognized agents
Apple, Meta, Amazon, Mistral, You.com, Common Crawl, Cohere, and others on the maintained allowlist
Onboarding and compatibility
Domain setup: Add the domain, prove ownership, configure the origin, and follow the dashboard’s DNS instructions.
Origin requirement: The origin must return meaningful HTML. SigilEdge is not a browser-rendering or SPA prerendering service.
Failure behavior: If enrichment is absent or generation fails, SigilEdge falls back to origin content instead of serving fabricated output.
deployment checklist
- 01Verify domain ownership
- 02Confirm the origin returns useful HTML
- 03Preview generated enrichment and set review thresholds
- 04Complete DNS routing and monitor coverage
Evaluate your site with the real workflow
Apply for early access or open the dashboard if you already have an account.