> For the complete documentation index, see [llms.txt](https://docs.seegov.org/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.seegov.org/api/data-and-errors.md).

# Missing data, notices and retries

`granularity=phrase` is the default. `granularity=word` adds `wordTimingAvailability: available|unavailable`. With no stored word timings, every phrase is still returned and each `words` value is null. Mixed or malformed word timings give 503 `data_unavailable`; phrase mode is independent of word timing validity. A successful response never silently drops or truncates phrases.

The API returns the original transcript text and timing. To recover analysis blocked by automated content filters, SeeGov may use a modified copy of selected passages for AI processing, with some text omitted or replaced. This can allow analysis of the rest of the meeting, but generated moment summaries and agenda analysis may omit details or lose context from those passages. These changes do not alter the transcript returned by the API. Processing notices identify portions where analysis could not be completed.

Speaker IDs identify speakers within a meeting and may reflect attribution corrections. Unknown names/types are null. Detail `speakersAvailability` can be available, partial, or unavailable, depending on known names.

Agenda responses distinguish unavailable analysis (`unavailable`, empty data) from an available result containing no agenda items (`available`, empty data). Numbering starts at one in response order and may differ from official agenda numbering. Descriptions retain completeness notices, including zero-duration notices about truncated analysis. `speakerIds` and `topics` are currently null.

Moment responses contain summaries and entries with `kind: processing_notice`, which describe analysis limitations for the indicated time ranges. Notice-only results are `available`; empty results are `unavailable` and do not distinguish no findings from analysis that is not available. An `available` response does not guarantee complete analysis. Tags are descriptive labels, not standardized topic classifications. Moment and word availability never use a partial state.

Errors use `application/problem+json`, with `status`, `code`, safe `detail` and `requestId`. Successful and error responses include `X-Request-ID` and `Cache-Control: private, no-store`. Larger responses negotiate gzip; use an HTTP client that decompresses it. HEAD uses the same access/limits but returns no body.

| Status                  | Action                                                                                                                         |
| ----------------------- | ------------------------------------------------------------------------------------------------------------------------------ |
| 400                     | Fix unknown/repeated/invalid options. For `invalid_cursor`, restart discovery from the last completed checkpoint.              |
| 401                     | Check or rotate the organization credential; do not retry indefinitely.                                                        |
| 403                     | Check service status/coverage with SeeGov; see [access guidance](https://docs.seegov.org/api/keys-and-access.md).              |
| 404                     | The requested resource does not exist.                                                                                         |
| 405                     | Use GET or HEAD.                                                                                                               |
| 409 `meeting_not_ready` | Retry the known ID later; it may be reprocessing.                                                                              |
| 409 `access_changed`    | Refresh coverage and rescan history with deduplication.                                                                        |
| 429                     | Honor `Retry-After` seconds, then back off with jitter.                                                                        |
| 500/503                 | Retry with bounded exponential backoff and jitter; honor `Retry-After`. Escalate repeated `data_unavailable` with request IDs. |

Starting organization defaults are 120 authenticated requests per UTC minute, including at most 30 transcript requests. All keys share those limits; retries, HEAD and denied attempts count. These limits are fixed rather than configured per organization. There is no monthly/lifetime transcript quota. Avoid synchronized minute bursts.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.seegov.org/api/data-and-errors.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
