Kagi Search API Personalization: Make Retrieval Tests Reproducible
Kagi API queries can inherit account lenses, upranks, downranks, and blocklists. Treat the account profile as part of every retrieval test.
From this desk
Start with the most useful page
The featured page offers a strong introduction. The remaining cards help readers move through this editorial section without turning the archive into a service menu.
Read the featured guide
Build an evidence-led publishing workflow
Move from reader job and claim classification through research, drafting, skeptical review, technical preparation, publication, and maintenance.
Field guide
Understand the reasoning
Use a guide when the task needs context, sequence, evidence standards, tradeoffs, and a defined final gate.
Checklist
Execute the final checks
Use the interactive pre-publish checklist when the method is understood and the work needs a repeatable release gate.
Browser tool
Preview one decision quickly
Use a focused tool when the task can be reduced to clear inputs, an immediate output, and an explicit limitation.
Complete guides
Work through the full method
Guides are reserved for tasks that need a substantial ordered workflow. Shorter decisions remain focused articles, checklists, or tools.
-
-
ChatGPT Restaurant Results and Reserve Buttons: A Verification Guide
Verify restaurant entity matching, location context, third-party provider inventory, Reserve-button handoff, and booking support boundaries in ChatGPT Search.
-
ChatGPT Shopping Labels and Review Summaries: What Merchants Can Control
Separate merchant-controlled product facts from ChatGPT-generated titles, feature labels, review summaries, and merchant ordering.
-
Control Google AI Previews with nosnippet, max-snippet, and data-nosnippet
Map Google’s nosnippet, max-snippet, and data-nosnippet controls to their documented effects across Search, AI Overviews, and AI Mode.
-
Googlebot’s 2 MB Fetch Limit: Audit What Appears Before the Cutoff
Googlebot fetches only the first 2 MB of supported non-PDF files. Audit the byte order of titles, canonicals, main content, links, and structured data.
-
Perplexity Agent API Citations: Normalize the Source Contract Before Migrating
Normalize numbered and source-typed Agent API citations into one internal source contract before migrating from legacy Sonar or MCP-backed workflows.
-
Google “Read More” Deep Links: Preserve Section Anchors and Hash Fragments
Keep target answers visible, preserve hash fragments, and stop page-load scripts from overriding section anchors used by readers and Google snippets.
-
Web Bot Auth for Publishers: Verify Signed AI Agents Without Breaking Crawlers
Test signed AI-agent verification at the CDN, WAF, or origin while preserving IP, reverse-DNS, and user-agent fallback for unsigned requests.
-
Back Button Hijacking: Audit Third-Party Scripts Before Google Treats It as Spam
Reproduce browser-history manipulation across real entry, consent, ad, and device states, then isolate the first- or third-party script that causes it.
A useful standard
What qualifies as a SearchEngineAnswer guide
A guide earns its length by helping a reader complete a demanding task. It should be navigable, testable, and useful after the first read.
A bounded reader job
The opening states who the guide is for, what decision or task it supports, and what remains outside its scope.
An ordered method
Sections follow the work itself. Examples, evidence, and limitations appear beside the decisions they constrain.
A completion gate
The ending defines what finished looks like, what still needs monitoring, and which tool or checklist supports the next step.
References and next steps