Committees¶

Committee work is exposed two ways:

GET /committees/meetings — distinct meetings list, derived from Hansard speech rows. Free-tier; no TEI cost.
/search/speeches?speech_type=committee — full-text search inside committee transcripts. Pro-tier (TEI-embedded; see Rate limiting).

Coverage as of v1.0:

Jurisdiction	Source	Speeches	Notes
Federal (HoC committees)	openparliament.ca	~5.5K (P45-S1)	100% MP-FK
Alberta (standing committees)	assembly.ab.ca PDFs	~10K (L31-S2)	First provincial slice; 31% MLA-FK (transcripts are ~40% witnesses)
BC	hansard-bc	~34K	Mostly Committee of the Whole — sittings in committee form, not standing-committee transcripts. Real BC standing-committee pipeline is still on the roadmap.

Other provinces don't have committee transcripts ingested yet.

`GET /committees/meetings`¶

Distinct committee meetings inferred from speeches with speech_type='committee'.

Query parameters¶

Name	Type	Notes
`level`	`federal` \| `provincial` \| `municipal`	Filter by chamber level.
`province_territory`	2-letter code	e.g. `AB`, `BC`. Federal: NULL.
`source_system`	string	Direct upstream filter — `openparliament` (federal), `assembly.ab.ca` (AB), `hansard-bc` (BC).
`from`	`YYYY-MM-DD`	Lower bound on first spoken-at within meeting.
`to`	`YYYY-MM-DD`	Upper bound.
`page`	int ≥ 1	Default `1`.
`limit`	int 1–100	Default `50`.

Response¶

{
  "items": [
    {
      "meeting_url": "https://openparliament.ca/committees/justice/45-1/29/",
      "level": "federal",
      "province_territory": null,
      "source_system": "openparliament",
      "date": "2026-05-06",
      "first_spoken_at": "2026-05-06T15:30:00Z",
      "last_spoken_at": "2026-05-06T17:25:00Z",
      "speech_count": 184
    },
    {
      "meeting_url": "https://docs.assembly.ab.ca/LADDAR_files/docs/committees/hs/legislature_31/session_2/20260413_1030_01_hs.pdf",
      "level": "provincial",
      "province_territory": "AB",
      "source_system": "assembly.ab.ca",
      "date": "2026-04-13",
      "first_spoken_at": "2026-04-13T16:30:00Z",
      "last_spoken_at": "2026-04-13T18:15:00Z",
      "speech_count": 142
    }
  ],
  "page": 1,
  "limit": 50,
  "total": 5766,
  "pages": 116
}

How `meeting_url` is computed¶

There's no committee_meetings table — meetings are derived per-row.

Federal openparliament.ca: source URL is per-speech-turn (https://openparliament.ca/committees/justice/45-1/29/larry-brock-1/). The meeting key strips the per-speaker slug: https://openparliament.ca/committees/justice/45-1/29/. The path segment after /committees/ is the committee acronym; the next is <parliament>-<session>; then the meeting number.
AB assembly.ab.ca: source URL is the meeting PDF; the meeting key is the URL itself. Committee acronym is the path segment after /committees/ (e.g. hs for Health).

If your client needs the committee acronym specifically, parse it out of meeting_url — exposing a committee_acronym field as a first-class column is a v1.1 candidate.

Example¶

curl -s -H "Authorization: Bearer cpd_live_…" \
  "https://canadianpoliticaldata.org/api/public/v1/committees/meetings?level=federal&from=2026-01-01" \
  | jq '.items[] | {date, meeting_url, speech_count}'

Full-text search inside transcripts¶

For "find every committee speech mentioning X" use the existing search endpoint with the speech_type=committee filter:

curl -s -H "Authorization: Bearer cpd_live_…" \
  "https://canadianpoliticaldata.org/api/public/v1/search/speeches?q=carbon%20capture&speech_type=committee&limit=20"

This is the PRO-tier search route — it routes the query through TEI for embedding and shares the 8-slot concurrency semaphore with all other paid search calls. See Rate limiting for the 503 + Retry-After semantics.

Recipes¶

Walk every meeting for one federal committee in a parliament¶

The meeting URL convention means you can filter post-fetch by URL substring:

curl -s -H "Authorization: Bearer cpd_live_…" \
  "https://canadianpoliticaldata.org/api/public/v1/committees/meetings?level=federal&limit=100" \
  | jq '.items[] | select(.meeting_url | contains("/justice/45-1/"))'

Fetch every speech in one meeting¶

The meeting_url is a prefix; combine with the speech search:

MURL="https://openparliament.ca/committees/justice/45-1/29/"
curl -s -H "Authorization: Bearer cpd_live_…" \
  "https://canadianpoliticaldata.org/api/public/v1/search/speeches?q=&speech_type=committee&limit=50" \
  | jq --arg m "$MURL" '[.items[] | select(.source_url | startswith($m))]'

(The empty q= triggers the route's session-listing fallback mode; substitute a real query term for actual relevance ranking.)

Caveats¶

BC misclassification. A large chunk of BC speech_type='committee' rows are sittings of the Committee of the Whole inside the BC Legislative Assembly chamber — NOT standing-committee transcripts. When BC's real provincial-committee pipeline ships, these may be reclassified.
MLA-FK rates are lower than chamber Hansard. Committee meetings bring in witnesses (not politicians), so the unresolved speeches in a committee transcript are usually correctly-unresolved witnesses, not a parsing failure.
No committee_acronym or committee_name as first-class fields in v1.0. Parse them out of meeting_url. Promotion to columns is a v1.1 candidate once the column-derivation logic stabilises across more provinces.