Congress Search

Search congressional hearings, floor speeches, and votes using the PolSearch API.

Triggers

Use this skill when the user wants to:

•Search congressional hearings
•Find floor speeches by topic, speaker, or date
•Look up what was said in Congress about a topic
•Find statements by specific legislators
•Search for committee hearings
•Query the congressional record

API Base URL

code

${BASE_URL}

Set BASE_URL to the API server address (e.g., http://localhost:3000 or your deployed server).

Quick Start

Basic Search

Search for mentions of "climate change" in congressional records:

bash

curl "${BASE_URL}/search?q=climate%20change"

Search Floor Speeches

Find floor speeches about immigration:

bash

curl "${BASE_URL}/search?q=immigration&type=floor_speech"

Search by Speaker

Find statements by a specific speaker:

bash

curl "${BASE_URL}/search?q=healthcare&speaker=Pelosi"

Search Hearings by Committee

Search for hearings from a specific committee:

bash

curl "${BASE_URL}/search?q=cybersecurity&type=hearing&committee=Homeland%20Security"

Search with Date Range

Find content from a specific time period:

bash

curl "${BASE_URL}/search?q=infrastructure&from=2023-01-01&to=2023-12-31"

Search by Chamber

Search only House or Senate content:

bash

curl "${BASE_URL}/search?q=budget&chamber=senate"

Search Votes

Find roll call votes on a topic:

bash

curl "${BASE_URL}/search?q=immigration&type=vote"

Get Vote Details

Get full vote information including vote counts:

bash

curl "${BASE_URL}/content/{vote-uuid}"

Returns vote result, vote type, category, and full breakdown (yea/nay/present/not_voting).

Get More Context

Include surrounding segments for better context (up to 10 segments before/after):

bash

curl "${BASE_URL}/search?q=tariffs&context=3"

Get Full Content Details

After finding a result, get full metadata by content ID:

bash

curl "${BASE_URL}/content/{uuid}"

Search Modes

Mode	Description	Use Case
`hybrid`	Combines vector + full-text search (default)	Best overall quality
`vector`	Semantic similarity using embeddings	Finding conceptually related content
`fts`	Keyword-based full-text search	Finding exact terms
`phrase`	Exact phrase matching	Finding specific quotes

Example with mode:

bash

curl "${BASE_URL}/search?q=tax%20reform&mode=phrase"

Content Types

Type	Description
`hearing`	Congressional committee hearings
`floor_speech`	Speeches on the House or Senate floor
`vote`	Voting records
`all`	Search all content types (default)

Multiple types can be comma-separated:

bash

curl "${BASE_URL}/search?q=defense&type=hearing,floor_speech"

Response Format

Search results include:

•content_id: UUID of the hearing/speech
•text: The matching text segment
•score: Relevance score (0-1, higher is better)
•content_type: Type of content (hearing, floor_speech, vote)
•speaker_name: Who said it (if available)
•title: Title of the hearing/speech
•date: Date of the content
•source_url: Link to original document on GovInfo
•committee: Committee name (hearings only)
•chamber: House, Senate, or both
•congress: Congress number (hearings only)
•context_before/context_after: Surrounding segments (when context > 0)

Pagination

Use limit and offset for pagination:

bash

# First page (10 results)
curl "${BASE_URL}/search?q=education&limit=10"

# Second page
curl "${BASE_URL}/search?q=education&limit=10&offset=10"

The response includes has_more and next_offset to facilitate pagination.

Full API Reference

See references/api_reference.md for complete endpoint documentation, parameter details, and response schemas.