About
A bird encyclopedia that shows its work.
KnowThePerch is built on one rule: if a fact can be looked up from a database, it was looked up — not generated by AI. Every number on this site traces back to eBird, IUCN, or Xeno-canto.
A field atlas, not a content mill.
KnowThePerch is a bird encyclopedia covering North American species. It’s built as a programmatic site — meaning thousands of pages are generated from structured data rather than written one by one.
The site currently targets 10,000–25,000 pages across species profiles, bird families, state-level species lists, side-by-side comparisons, and practical birding guides.
What makes KTP different from other bird sites is the separation between data and prose. Measurements, taxonomy, range maps, frequency charts, and conservation status are never AI-generated — they come from APIs. Only the descriptive text explaining those facts is written with AI assistance, and it’s reviewed before publication.
Data in, prose out, review before publish.
Stage 1 — Data collection
Taxonomy, measurements, range, and conservation status are pulled from eBird (Cornell Lab), IUCN Red List, and Xeno-canto. Observation frequency comes from eBird bar chart data — real checklists from real birders, not estimates.
Stage 2 — Prose generation
Descriptive text is generated by Qwen3, a locally-hosted language model. It receives structured data and writes explanations. It cannot invent numbers, manufacture sightings, or contradict the injected data. Generated text is cached in SQLite and reviewed before publication.
Stage 3 — Template rendering
Data and prose are assembled into PAGE_DATA objects and rendered by WordPress templates. The templates decide layout, navigation, and what appears on each page type — species profiles, comparisons, state pages, and guides each have their own template.
Where the numbers come from.
Taxonomy, species codes, common and scientific names. Observation frequency from bar chart API — the percentage of checklists reporting a species in a given region and month. Updated annually with eBird taxonomy revisions.
Conservation status categories: Least Concern (LC), Near Threatened (NT), Vulnerable (VU), Endangered (EN), Critically Endangered (CR). Updated when IUCN publishes revised assessments.
Bird call and song recordings. Embedded audio players on species profiles link directly to CC-licensed recordings from Xeno-canto’s community-contributed database.
All API data is stored locally in SQLite. This is the single source of truth for every number on the site. The pipeline (n8n) pulls from APIs → writes to SQLite → generates prose → publishes to WordPress.
What AI writes and what it doesn’t.
Species profiles, state guides, and comparison pages use AI-assisted prose. The descriptive text on those pages was generated by Qwen3 (a locally-hosted language model) based on verified data, then reviewed by a human editor before publication.
AI writes: descriptive paragraphs explaining bird behavior, habitat preferences, identification tips, and seasonal patterns.
AI never writes: measurements, species counts, conservation status, frequency percentages, hotspot names, taxonomy, or any numerical claim. Those come from APIs.
Editorial content (blog posts, birding guides) is written by human authors without AI generation.
Found an error? Tell us.
If you find incorrect information on any page — a wrong measurement, outdated conservation status, or a prose claim that contradicts the data — contact us with the page URL and the specific error. We review correction requests within 5 business days.
Confirmed errors are corrected at the source (SQLite database or prose template), not patched in the page. The page is then regenerated and the updated date is revised.
Report an error →