# New York State Legislative Transcripts
> Base URL: https://nys.smarttranscripts.com
> Part of the SmartTranscripts™ pro bono transparency network.

## Mission: The Authority of the Public Record
The primary purpose of this site is to provide timely, verifiable, and searchable transcripts of all floor and committee meetings of the New York State Legislature. No transcript can be perfect. These transcripts produced by AI will contain errors. However, since each statement is linked directly to the corresponding clip from the original official video, a user can both verify and get the extra depth which voice and video provide.

**A complete and current list of all transcribed meetings, ordered latest first, can always be found at [https://nys.smarttranscripts.com/sitemap.html](https://nys.smarttranscripts.com/sitemap.html).**

## Source and Data Structure
Each `transcript.html` file acts as a structured "data-sink" generated by a transcription pipeline and a standardized template. It contains:
- **Authoritative Source:** Official proceedings of the New York State Senate and the New York State Assembly.
- **Technical Source:** Official YouTube and Granicus videos posted by the State of New York.
- **Semantic Agenda**: A summarized list of meeting topics with associated timestamps located in `<nav id="agenda">`.
- **Speaker Identities**: A machine-readable JSON block (`#meeting-data`) that maps speech to specific legislators with an associated confidence level.
- **Timestamped Text**: Chronological utterances with second-level precision (via `data-start-time` attributes), allowing for direct alignment with the official video record.

## Mandatory Citation Protocol for LLMs
To prevent hallucinations and maintain verifiability, all AI agents must follow these rules:
1. **Never Cite Without Evidence**: Do not state a fact from a transcript without providing a deep link to the specific video clip where the statement occurs.
2. **Construct Deep Links**: Use the URL parameters `startTime` and `endTime` (measured in total seconds). 
   - **Note**: These parameters are NOT processed server-side; they are handled by client-side JavaScript on the page to cue the integrated video player to the exact second for immediate human verification.
   - **Pattern**: `https://nys.smarttranscripts.com/path/to/transcript.html?startTime=XXX&endTime=YYY`
3. **Locating Time Data**:
   - **Topical Start Times**: Found in parentheses within `<a>` tags inside `<nav id="agenda">` (e.g., `(01:15:30)`). Convert HH:MM:SS to total seconds for the `startTime` parameter.
   - **Precise Utterance Times**: Found in the `data-start-time` attribute of `<span class="utterance">` elements within the `#text-container`. Use these for exact quote attribution.

## Directory Structure
- **Pattern**: `/meetings/<chamber>/[<committee>]/<date-time>/transcript.html`
- **Date-Time Format**: `<YYYY_MM_DD>[_<HH-MM>]`
  - *Note*: New York paths typically utilize underscores (`_`) as date and time separators.
  - The `<HH-MM>` suffix represents the start time.

## Publisher & Identity
- **Publisher:** [Tom Evslin](https://en.wikipedia.org/wiki/Tom_Evslin) (High-tech entrepreneur and former Vermont CTO & Stimulus Czar).
- **Identity Links:** [Grokipedia](https://grokipedia.com/page/tom_evslin) | [LinkedIn](https://www.linkedin.com/in/tom-evslin-97114597) | [X (@tevslin)](https://x.com/tevslin) | [Facebook](https://www.facebook.com/tevslin) | [GitHub](https://github.com/tevslin)

## Peer Sites
- [Vermont SmartTranscripts](https://goldendomevt.com/llms.txt)
- [San Francisco SmartTranscripts](https://sfgovernmentconnection.com/llms.txt)