SRAC Search API

The Social Rights Advocacy Centre maintains a searchable research collection on social and economic rights, Charter litigation, housing, poverty, equality and international human rights law. The same search that powers this site is available as an open API, so that researchers, reference tools and AI research agents can find and cite the original documents directly.

What is indexed

• Every HTML page on socialrights.ca (and the companion socialrightscura.ca archive).
• Every PDF the site publishes — factums, decisions, UN submissions, reports — indexed page by page, so a result points at the specific page that matches.
• OCR is applied to scanned PDFs where possible, so image-only documents are searchable too.
• The index is rebuilt nightly whenever the site changes. Every response carries a build block that says when it was built and how much it covers.

Endpoint

The API is served for SRAC from jacksaiapps.com. This is the official Social Rights Advocacy Centre search endpoint; there is no separate socialrights.ca host for it.

No authentication, no API key. Plain HTTPS GET; responses are JSON (application/json, UTF-8) and allow cross-origin requests from any page.

Example

curl "https://jacksaiapps.com/srac/public-api/site-search?q=housing&limit=2"

{
  "query": "housing",
  "results": [
    {
      "kind": "pdf",
      "title": "Volume 11 of Heegsma AR",
      "url": "https://www.socialrights.ca/Heegsma/Volume 11 of Heegsma AR.pdf#page=234",
      "rel_path": "Heegsma/Volume 11 of Heegsma AR.pdf",
      "page": 234,
      "snippet": "…the right to adequate <mark>housing</mark> under…",
      "snippet_text": "…the right to adequate housing under…",
      "score": -12.4
    }
  ],
  "scope": { "source_pages": [] },
  "build": {
    "built_at": "2026-08-11T07:09:26+00:00",
    "html_count": 141,
    "pdf_count": 1684,
    "row_count": 92027,
    "citation_count": 1434,
    "orphan_count": 755,
    "warning_count": 4
  }
}

Query parameters

Response fields

Results are one per document: all pages of a PDF collapse into its single best-matching page. They are ranked by relevance, with HTML pages ordered above PDFs at equal relevance.

Each result

Listing the collection

A second endpoint returns a paginated manifest of every indexed document — metadata only, no text — so a client can learn the scope of the collection without harvesting it.

The response has documents, total, offset, limit, next_offset (null on the last page), filters and the same build block. Ordering is by kind (pages first) then path, and is stable for a given index build. Each document has:

The date a document was indexed is the manifest's build.built_at; the index does not track per-document modification dates.

Rate limits and errors

The API is public and free to use for research, but it runs on modest hardware, so there are limits — set high enough that any reasonable research use will never meet them:

• 300 requests per minute and 3,000 per hour per client IP address;
• 1,200 requests per minute across all clients combined.

Every response includes X-RateLimit-Limit, X-RateLimit-Remaining and X-RateLimit-Reset headers. Over the limit you receive HTTP 429 with a JSON body ({"error": "rate_limit_exceeded", …}) and a Retry-After header — wait that many seconds and retry. Bulk copying of the whole corpus is not what the API is for; if you need the collection for a legitimate project, please get in touch.

While the index is being rebuilt or is otherwise unavailable, the API returns HTTP 503 with an error field and empty results. This is brief and transient; retry shortly.

Machine-readable descriptions

• OpenAPI 3.1: socialrights.ca/openapi.json (also served live at jacksaiapps.com/srac/public-api/openapi.json).
• llms.txt: socialrights.ca/llms.txt — a short plain-text pointer for AI systems.
• Sitemap: socialrights.ca/sitemap.xml.

Terms of use and citation

This API is provided to facilitate research, retrieval, citation, and analysis of materials in the Social Rights Advocacy Centre collection. Please use reasonable request rates and cite original documents using the URLs provided in search results.

Search results are pointers, not the record: always cite and link the original document at its url. Snippets are short excerpts intended to help you find the right document, not a substitute for reading it. The API is read-only and stores no personal data beyond ordinary access logs.

Try it

Press Run to see a live response.