Japan Diet Parliament Debates & Speeches Search (国会会議録)

What it does

Search what was said in Japan's parliament, the National Diet (the Japanese equivalent of Hansard): plenary sessions and committees of the House of Representatives (衆議院) and the House of Councillors (参議院), from 1947. Results come as structured rows: date, house, meeting, speaker, party, position, the full text of the speech, and links to the official record.

Data source: the National Diet Library's official speech search API of the 国会会議録検索システム. No scraping of web pages.

What you can do with Diet minutes data

• Policy research: how often and how a topic (for example 子育て支援, child-care support) is discussed.
• Journalism: find what a minister or member said, with a link to the official record.
• Speaker analysis: compare statements by party, committee or period.

Why choose this Actor

• Uses the National Diet Library's official API, not page scraping.
• Low price: $2 per 1,000 speeches.
• Every row has source links (speechURL, meetingURL, pdfURL).

Quick start

Keep the example (子育て支援 = child-care support, 20 speeches) and press Start.

Input

{
  "keyword": "子育て支援",
  "maxResults": 20
}

Give at least one search condition. All conditions are combined with AND. To narrow the example, add for instance "nameOfHouse": "衆議院", "nameOfMeeting": "予算委員会", "dateFrom": "2024-01-01" and "dateUntil": "2024-12-31".

Field	API parameter	Notes
keyword	any	Several words separated by spaces must all match
speaker	speaker	e.g. 石破茂
speakerGroup	speakerGroup	Parliamentary group (会派)
speakerPosition	speakerPosition	e.g. 内閣総理大臣
nameOfHouse	nameOfHouse	衆議院 / 参議院 / 両院 / 両院協議会. Other values are rejected; the API would silently ignore them and widen the search.
nameOfMeeting	nameOfMeeting	e.g. 本会議, 予算委員会
dateFrom, dateUntil	from, until	YYYY-MM-DD
sessionFrom, sessionTo	sessionFrom, sessionTo	Diet session numbers (国会回次)
searchRange	searchRange	冒頭 / 本文 / 冒頭・本文
maxResults	—	1–10,000, default 100
includeMeetingHeader	—	Include speechOrder 0, the 会議録情報 header (attendance, agenda). Default off.
excludeWitnesses	—	Drop speeches by 証人・参考人・公述人. Default off.

Output

{
  "speechID": "121405254X00220241004_016",
  "issueID": "121405254X00220241004",
  "session": 214,
  "nameOfHouse": "衆議院",
  "nameOfMeeting": "本会議",
  "issue": "第2号",
  "date": "2024-10-04",
  "closing": null,
  "speechOrder": 16,
  "speaker": "石破茂",
  "speakerYomi": "いしばしげる",
  "speakerGroup": "自由民主党・無所属の会",
  "speakerPosition": "内閣総理大臣",
  "speakerRole": null,
  "speech": "○内閣総理大臣（石破茂君）　…",
  "startPage": 2,
  "speechURL": "https://kokkai.ndl.go.jp/txt/121405254X00220241004/16",
  "meetingURL": "https://kokkai.ndl.go.jp/txt/121405254X00220241004",
  "pdfURL": "https://kokkai.ndl.go.jp/img/121405254X00220241004/2",
  "source": "国会会議録検索システム（国立国会図書館） https://kokkai.ndl.go.jp/"
}

Results are ordered by meeting date, newest first. The API fixes this order and offers no other sort. The status message reports how many speeches matched in total.

Pricing (pay per event)

Event	Price	When
speech-record	$0.002 ($2 per 1,000 speeches)	Each speech saved
Actor start	Apify default	Once per run

If you set a maximum charge for the run, the Actor stops once it is reached.

FAQ

Which language are the speeches in? Japanese. Field names and this README are English; keywords should be Japanese.

Is it complete? It returns what the National Diet Library's search API returns, from 1947. The status message reports how many speeches matched in total.

Can I use the text commercially? See the copyright note below and check it for your own use.

Responsible use (please read)

• Polite access. The service asks automated clients not to send parallel requests and to wait a few seconds after each response. This Actor sends one request at a time, 100 speeches per page, and waits 3 seconds between pages. Large searches take a while: 10,000 speeches need about 100 requests, roughly 5 minutes.
• robots.txt at kokkai.ndl.go.jp is checked before the first request. If it disallows the API path, or cannot be fetched, the run stops without searching.
• Attribution. Every row carries a source field. When you publish results, credit the 国会会議録検索システム (National Diet Library) and link to speechURL.
• Copyright. Under Article 40(1) of Japan's Copyright Act, political statements made in public in the Diet may be used freely, except compiling one person's statements into a collection. Check this for your own use; this is not legal advice.
• Personal data. Speakers are mostly legislators and officials acting in public roles. Witnesses and invited speakers (証人・参考人・公述人, speakerRole set) can be private individuals. Use excludeWitnesses if you do not need them, and do not use the data to profile individuals.
• Errors from the API, such as an invalid date, are shown in the run's status message. Timeouts, HTTP 429 and 5xx are retried 3 times with backoff.

Run locally

python -m venv .venv && . .venv/bin/activate
pip install -r requirements.txt
python -m unittest discover -s tests -t .      # unit tests, API mocked

mkdir -p storage/key_value_stores/default
echo '{"keyword": "子育て支援", "maxResults": 20}' > storage/key_value_stores/default/INPUT.json
python -m src                                   # or `apify run`; results in storage/datasets/default/
ACTOR_TEST_PAY_PER_EVENT=true python -m src     # simulate pay-per-event charging locally

Sources

• API documentation: https://kokkai.ndl.go.jp/api.html
• Help and terms: https://kokkai.ndl.go.jp/help.html

Related actors

• AI Crawler Checker & GEO / AEO Audit: check a site for AI search readiness (llms.txt, AI bots, JSON-LD).
• Video Frame Extractor: frames, scene changes and thumbnails from video file URLs.
• Stable Diffusion Prompt Extractor: read prompts and settings embedded in AI images.
• Japanese to Romaji, Hiragana & Furigana Converter: readings for Japanese text.