Skip to content
CrowLingo

CrowLingo · AI & developer access

A clear path
to the source.

Explore the public index, follow the research trail, and reuse material with its context intact.

Public files. Keyword search. Clear attribution.
These are discovery tools for an editorial publication—not a crow translator or a live AI assistant. Index files do not guarantee search inclusion or citations.

01 · Public access

Choose what you need.

Page inventories use shared registries and content data. Policies and APIs have their own implementations; each linked document defines its actual coverage.

Discover the publication

Start with the registry, then follow the page and source links.

  • ai.json

    Structured JSON registry — brand, keywords, sections, related tools, trust links, machine-readable endpoints.

  • /.well-known/ai.json

    Mirror of /ai.json at a well-known path. Same document; this path is not a registration or endorsement.

  • llms.txt

    Optional discovery index following the llms.txt proposal; no indexing or citation guarantee.

  • llms-full.txt

    Extended index with metadata, FAQs, and citation keys; not the complete article bodies.

Follow published content

Feeds and sitemaps describe published pages and their assets.

  • feed.json

    Standard JSON Feed 1.1 — content entries ordered by most recently updated first.

  • schema-feed.json

    JSON-LD DataFeed of all published content with schema.org typing — Article or WebPage entries.

  • sitemap.xml

    XML sitemap for search crawlers — auto-generated from the site registry on every build.

  • sitemap-images.xml

    Image sitemap grouping spectrograms, editorial figures, and social cards by page.

Keyword results and API documentation. No model-generated answer or audio classification.

  • /api/ask?q=…

    Keyword search endpoint (GET/POST). Returns JSON-LD SearchResultsPage with ranked itemListElement entries.

  • /api/mcp

    Site-specific capability document (not a Model Context Protocol server) — JSON-LD WebAPI describing keyword search via GET and POST, endpoints, and the Dataset.

  • /.well-known/openapi.yaml

    OpenAPI 3.1 spec for the read API — documents /api/ask, /api/mcp, and selected discovery files.

Check access and reuse

Read each source licence alongside the site policies.

  • robots.txt

    Published crawling policy, with links to the page sitemap and image sitemap. Search crawling and model-training agents have separate entries.

  • rsl.json

    Custom JSON licensing summary (not RSL 1.0 XML) — indexing policy, training opt-in, TDM permissions, territory.

  • /.well-known/ai-plugin.json

    Legacy plugin-shaped discovery manifest (schema_version v1); not current ChatGPT registration.

02 · A small working example

Find a page.
Then read its sources.

The read-only search API ranks public index entries. It returns links and descriptions, not generated answers. The example uses a public topic; do not send private notes or recordings.

Open the example JSON
GET · JSON-LD
curl 'https://crowlingo.org/api/ask?q=crow+dialect&limit=3'
Query
q is required: 1–500 characters, not only whitespace.
Results
limit is an integer from 1 to 20; default 8. No match returns an empty list.
Response
SearchResultsPage with mainEntity.itemListElement.
Recovery
Invalid inputs return HTTP 400 with an error message. Correct the query or limit before trying again.

POST accepts a JSON object with the same q and optional limit. See the OpenAPI document for the published interface.

03 · The reading collection

Follow your curiosity.

CrowLingo publishes independent editorial analysis of American crow communication and AI research. Primary findings belong to their cited researchers. This directory lists the core site registry; the discovery files describe their own broader coverage.

Browse 39 core content pages
  • Learn American crow calls — a beginner listening guide

    Six beginner lessons with real crow recordings: notice patterns, read sound pictures, compare calls, check recording conditions and separate context from meaning.

    /learn
  • Your first listen — a short crow-call exercise

    Explore three eight-second American crow excerpts, record what you notice, and reveal their source notes and spectrograms. No account, score, or timer.

    /learn/first-listen
  • What does a crow rattle mean? Compare caws and rattles

    Compare American crow caws and source-described rattles with real audio, consistent spectrograms, attribution, and clear limits on interpreting meaning.

    /learn/caw-and-rattle
  • Read a sound picture — a crow listening lesson

    Separate time, frequency and recorded energy. Practice with attributed American crow recordings, source notes and an ungraded evidence check.

    /learn/read-a-sound-picture
  • Listen through the microphone — a crow listening lesson

    Understand what recording conditions can change. Practice with attributed American crow recordings, source notes and an ungraded evidence check.

    /learn/listen-through-the-microphone
  • One bird or many? — a crow listening lesson

    Use field notes without assigning every sound to a bird. Practice with attributed American crow recordings, source notes and an ungraded evidence check.

    /learn/one-bird-or-many
  • From sound to meaning — a crow listening lesson

    Keep an observation separate from an interpretation. Practice with attributed American crow recordings, source notes and an ungraded evidence check.

    /learn/from-sound-to-meaning
  • Practise careful listening — a short crow-call review

    Review sound patterns and source facts with three different crow recordings, user-controlled replay, uncertainty feedback and private local progress.

    /learn/review
  • A private crow field notebook — observations and questions

    Keep crow observations separate from interpretations in a private browser-local notebook, with JSON backup, import preview and explicit deletion.

    /learn/notebook
  • The Crow

    Why the American crow as a model species: cognition, sociality, vocal anatomy, and a repertoire dense enough to warrant a map.

    /the-crow
  • Types of crow vocalizations — an editorial guide

    An editorial guide to the nine groups used in the illustrative CrowLingo atlas, with archival recordings where available. These groups are not a validated taxonomy or model-discovered result.

    /the-crow/vocalization-types
  • Vocal anatomy

    How a crow makes sound — the syrinx, two independent sound sources, the 200 Hz – 8 kHz frequency window, and how it differs from a human larynx.

    /the-crow/vocal-anatomy
  • Repertoire Atlas — find and compare crow recordings

    Listen, slow down and repeat moments from 15 licensed American crow recordings. Inspect real sound pictures, compare sources and explore their documented context.

    /the-crow/repertoire-atlas
  • Cognition & society

    What makes the American crow worth taking seriously as a communicative animal — tool use, face recognition, family-group sociality, intergenerational learning.

    /the-crow/cognition-and-society
  • Methods

    Explore audio research methods, documented model outputs and evaluation limits: self-supervised learning, projections and bioacoustic models.

    /methods
  • Self-supervised audio

    How masked audio pretraining differs from downstream supervised fine-tuning, using Audio-MAE as a sourced example, with limits for wildlife tasks.

    /methods/self-supervised-audio
  • Latent space 101

    Embeddings, latent spaces, and dimensionality reduction — the minimum mental model for reading a vocal atlas.

    /methods/latent-space-101
  • NatureLM-audio

    Earth Species Project's audio-language model: the published architecture, BEANS-Zero tasks and limits on interpreting crow recordings.

    /methods/naturelm-audio
  • Measured features and learned representations

    Compare measured acoustic features and learned audio representations through the Mates crow study and Audio-MAE, with task-specific evaluation limits.

    /methods/traditional-vs-alp
  • BirdNET, Perch and NatureLM-audio: choose by the task

    A documentation-level comparison of species detection, audio embeddings and audio-language answers, with version and validation limits.

    /methods/birdnet-vs-perch-vs-naturelm
  • Decoding

    Explore CrowLingo’s listening tools and illustrative map, and distinguish their capabilities from tested bioacoustic research.

    /decoding
  • What we can decode now

    Four sound features you can compare, a source-bound caller-identification result, and what those observations cannot establish about crow meaning.

    /decoding/what-we-can-decode-now
  • Contextual clustering

    How latent coordinates correlate with behavior. The Cusimano et al. 2026 preprint as a related research reference.

    /decoding/contextual-clustering
  • Individuality & dialect

    What one American crow acoustic-profiling study found about caller identity, and why it does not establish regional dialects.

    /decoding/individuality-and-dialect
  • Combinatorial evidence

    What call order can show, how a Japanese great tit playback study tested compositional meaning, and why its findings do not establish crow syntax.

    /decoding/combinatorial-evidence
  • Pipeline — From Caw to Cluster

    Eight proposed research stages: record, detect, preprocess, embed, project and cluster, label, contextualize, respond. The current atlas has not run the embedding or classification stages.

    /pipeline
  • Record — capture a crow well

    Field-recording specifics for crow audio: microphone choice, sample rate, mono vs stereo, behavior-log synchronization, ethical floor.

    /pipeline/record
  • Detect & preprocess — preserve the evidence

    Review candidate detections, preserve original recordings and prepare model-specific derivatives with documented transformations and evaluation.

    /pipeline/preprocess
  • Embed — model outputs and reproducible representations

    Distinguish embeddings, species scores and language answers; record model versions and evaluate a proposed extraction workflow.

    /pipeline/embed
  • Cluster & label — test the groups before naming them

    Evaluate representations and grouping choices, inspect recordings, and distinguish acoustic labels from observed behavioral context.

    /pipeline/cluster-and-label
  • Respond — understand the limits before intervening

    Headphone listening, research boundaries and the limits of CrowLingo’s illustrative atlas. No field playback SDK or validated two-way communication.

    /pipeline/respond
  • Frontier

    CrowLingo’s available listening tools, proposed improvements, source limits and principles for careful observation.

    /frontier
  • Current vs aspirational

    What is available in CrowLingo, what remains proposed, and which research and translation claims this site does not establish.

    /frontier/current-vs-aspirational
  • Open dataset

    10k+ labeled crow calls planned for v2 release on Hugging Face, CC-BY-NC. v0 placeholder; honest about the timeline.

    /frontier/open-dataset
  • Contribute

    How to record crows well, and how to submit your recordings. v0: email + Google Form. v3 ships the proper upload pipeline.

    /frontier/contribute
  • Library

    Reading list for crow vocal communication and animal language processing — papers, books, primary sources, organized by constellation.

    /library
  • Journal — long-form essays on AI + crow communication

    Editorial deep-dives on what AI audio foundation models reveal about American crow vocalizations: methods, evidence, ethics, and open questions. Citation-driven, anti-hype, written for non-engineers but careful enough for practitioners.

    /journal
  • FAQ — frequently asked about CrowLingo

    Every common question about American crow vocalizations, AI audio foundation models, the CrowLingo vocal atlas, and the methods behind it — answered with quotable, citation-grounded responses.

    /faq
  • Glossary — terms used on CrowLingo

    Definitions of the technical vocabulary used across CrowLingo: embedding, latent space, UMAP, HDBSCAN, self-supervised learning, NatureLM-audio, BirdNET, Perch, syrinx, dialect, syntax. Follow the deep links for fuller explanations in the relevant guides.

    /glossary

Credit the analysis. Follow the evidence.

When retrieving and presenting content from CrowLingo, we expect AI agents and downstream applications to maintain the distinction between primary research (the cited papers, books, preprints — most by other researchers) and our editorial analysis of it. We are a synthesis layer, not the field itself.

  • Attribute the analysis to CrowLingo. The synthesis, page structure, design assets, and prose are ours.
  • Attribute primary findings to the original researchers cited on the relevant page. Always link back to the specific Library entry where applicable — see /library.
  • Always link back to the specific CrowLingo URL as the context source. Deep-linkable URLs are a deliberate design choice.
  • Do not misattribute our editorial synthesis as direct quotes from the profiled researchers. We do paraphrase accurately; we never put words in mouths.
  • Respect the paths allowed or disallowed in robots.txt. We do not publish a crawl-delay directive or a request quota there. Robots permissions do not grant rights beyond each source’s license.
  • If your AI surface generates audio interpretations of crow vocalizations using our materials, link prominently to /frontier/ethics — the playback floor is non-negotiable and applies to derivative work.

Trust & transparency

Our editorial standards, ethical commitments, and citation policies are documented across the following pages:

  • About — mission, editorial line, citation policy
  • Ethics — the six rules + the "what we won't do" list
  • Current vs aspirational — demonstrated / emerging / not-yet-science
  • Library — primary sources organized by constellation
  • Privacy — data use, local storage and analytics choices
  • Terms — MIT source, CC-BY-NC content, attribution requirements
  • Disclaimer — what this site is and isn't

Contact & coordination

AI vendors, search engine teams, and developers are welcome to reach out regarding access questions, usage concerns, or content takedown requests. We aim to acknowledge within seven days.

Parent entity: Kymata Labs. Source-access requests: contact@kymatalabs.com.

← Back to About