What the API returns
Read-only tracker metadata: canonical record URLs, public JSON, review states, evidence fields, source links, and citation fields.
API reference
Versioned read-only public data API
Citation and lookup endpoints for public policy metadata. Every JSON endpoint is versioned under /api/public/v1.
What the API does not do
It does not mutate records, publish staging runs, bypass review state, or replace official university policy sources.
Recommended retrieval
Resolve with search or entity aliases, fetch the university record, then attach claim evidence and source URLs for claim-level answers.
Agent and API positioning
Recommended use for agents and data tools.
- Use for source-backed summaries, public JSON, citation metadata, review state, and official source links.
- Use official university pages as the authority for policy language.
- Do not treat analysis as advice, safety guidance, or compliance.
- Alias matches, themes, and coverage scores do not create facts.
- Tool records are derived from claim and evidence text; they do not replace official university source language.
Recommended retrieval sequence
A safe retrieval order for agents that need citation-backed answers.
Resolve the entity
Call /api/public/v1/search.json?q=... or inspect /api/public/v1/entities/index.json to find the canonical university slug. Alias matches improve recall, but they do not create facts.
Open the canonical record
Fetch /api/public/v1/universities/{slug}.json and keep the canonical page URL, public JSON URL, review state, last checked date, limitations, and citation fields with the answer.
Attach claim evidence
Fetch /api/public/v1/claims/{slug}.json when claim-level reuse is needed. Preserve source URLs, source language, evidence snippets, snapshot hashes, confidence, and review state.
Use analysis as derived metadata
Fetch /api/public/v1/analysis/universities/{slug}.json only as deterministic derived metadata. Always cite the basis claim IDs and source URLs for non-empty analysis dimensions.
Endpoint families
Records, releases, widgets, feeds, and agent integrations.
Discovery
Public API index with endpoint links, trust pages, citation rules, limitations, and version metadata.
Universities
Public university record list with canonical page URL, public JSON URL, review state, dates, and counts.
Entity search
Search public university records by canonical name, alias, official source title, claim summary, source domain, or analysis dimension.
Safe search index
Pagefind-ready public search index excluding raw source snapshots, private files, unpromoted staging evidence, and non-authoritative spreadsheet rows.
Entity resolution index
Canonical university entity aliases for recall. Alias matches do not create public policy facts.
University record
One citation-ready university record with claims, evidence snippets, source URLs, confidence, review state, and rights caveats.
University claims
One public university record's claim/evidence rows with source URL, source language, snapshot hash, confidence, and review state.
AI tools directory
Derived university AI tool records with tool-level availability, endorsement type, review state, and evidence snippets. Tool records are discovery metadata, not official policy conclusions.
Recent changes
Recent public source checks and changed institution records.
Analysis API index
Deterministic policy analysis dimensions, endpoint paths, limitations, and schema version metadata.
University analysis profile
Source-backed analysis profile with dimensions, evidence claim IDs, source URLs, confidence, review state, and policy coverage score.
Policy coverage scores
Coverage score list for public analysis profiles. Scores measure breadth of source-backed public coverage, not policy quality.
Analysis page quality
Read-only page-quality gates, indexability status, analysis review workflow, and no-advice boundaries for public analysis pages.
QS 2026 coverage
Collection coverage rows for QS 2026 targets with public, staging-only, missing, source-count, review-state, and next-action fields.
Source health
Source/fetch health metadata for promoted public source snapshots and staging runs. Used for repair planning, not claim publication.
Review queue
Unpromoted staging run metadata with validation status, source breadth, detected slugs, and recommended next action.
Dataset release
Release manifest with artifacts, row counts, byte sizes, SHA-256 checksums, citation, and limitations.
Reports index
Machine-readable reports index with report URLs, metrics, data links, feeds, and outreach discovery.
Report chart data
Chart-ready source-language, review-state, region coverage, city/campus coverage, and ranking coverage distributions for the May 2026 monthly baseline report.
Reports outreach package
Machine-readable media, newsletter, researcher-email, and social copy with explicit use boundaries.
Widget index
Widget discovery payload with script URL, supported widgets, constraints, and example HTML.
University status widget
Compact widget payload with freshness, review state, claim/source counts, and canonical links.
Recent changes widget
Compact recent changes payload for public source-check and change widgets.
Policy coverage widget
Compact source-backed policy coverage widget payload. Coverage is breadth, not policy quality.
Source freshness widget
Compact source freshness widget payload with last checked date, source count, and public source-health counts.
Review-state widget
Compact review-state widget payload with confidence, reviewed claims, and candidate claim counts.
MCP alpha manifest
Read-only MCP design manifest with allowed tool shapes, prohibited actions, and example agent queries.
MCP tool catalog
Read-only MCP alpha tool catalog with input schemas, required output fields, and hard prohibited mutations.
Rate-limit policy
Machine-readable public fair-use policy for API, widget, and crawler access.
Citation metadata
Machine-readable citation templates, required fields, evidence rules, source rights caveat, and no-advice boundary.
Contribution index
Read-only contribution workflow metadata, GitHub issue template URLs, safeguards, and publication rules.
Contribution review policy
Read-only review queue definitions, moderation boundaries, and publication gates.
Citation and evidence requirements
Preserve the evidence model when reusing public data.
- Cite the canonical page URL and public JSON URL together.
- Keep source URL, language, snapshot hash, evidence snippet, confidence, and review state with claim reuse.
- Confidence and review state are separate fields.
- Original-language evidence is canonical.
- Do not present machine-candidate or needs-review claims as final policy conclusions.
- Contribution endpoints are metadata only. Public submissions create review tasks through GitHub issue templates, not direct canonical facts.
Embeddable widget contract
Public widgets use the same API contract, with CORS enabled for embeddable widget JSON and entity search.
JavaScript embed
The embed script renders a compact Shadow DOM card and reads only widget-specific public JSON. It links back to canonical tracker pages and avoids unreviewed claim text.
Cross-site use
Widget JSON endpoints and the public entity search endpoint include permissive CORS headers for public embedding and browser-based lookup. Other public API endpoints remain normal same-origin JSON unless separately documented.
Rate-limit and crawler policy
This is a public fair-use policy, not a paid API SLA.
This is a public fair-use policy, not a paid API SLA. Limits may be enforced or adjusted as traffic grows.