Read a transcript by the ID returned from search_transcripts.
Input
| Field | Type | Default | Description |
|---|---|---|---|
id |
UUID | required | Transcript ID |
format |
txt, srt, vtt |
txt |
Rendered output format |
offset |
integer | 0 |
Character offset in the rendered output |
maxChars |
integer | 2000 |
Page size from 1,000 to 90,000 characters |
includeSegments |
boolean | false |
Include complete source segments contained in the page |
Output
The tool’s text content contains the rendered transcript page. Its structured result contains:
| Field | Type | Description |
|---|---|---|
id |
UUID | Transcript ID |
displayName |
string | Transcript name |
language |
string or null | Detected language |
duration |
number or null | Duration in seconds |
category |
category or null | Typist content category |
topics |
string[] or null | Extracted topics |
locked |
boolean | Whether transcript access is locked |
offset |
integer | Start of this page in the rendered output |
nextOffset |
integer | Start of the next page, when one is available |
totalChars |
integer | Total characters in the rendered output |
truncated |
boolean | Whether more content is available |
segments |
segment[] | Timing and speaker data when explicitly requested |
Pagination
When truncated is true, call the tool again with the returned nextOffset. Offsets refer to the
rendered format, so do not reuse a TXT offset for SRT or VTT.
The tool keeps transcript segments and subtitle cues together where possible.