Overview
Chapters divide audio recordings and transcripts into meaningful sections, making it easier to navigate, index, and reference specific parts of an event. Both the audio and transcript endpoints expose chapters through dedicated sub-endpoints. Each chapter includes a title, start timestamp, and end timestamp defining a distinct section within the content.Levels and nesting
Thelevel field defines the hierarchical depth of a chapter.
- Level 1: Top-level sections (e.g., “Prepared Remarks”, “Q&A”).
- Level 2: Subsections nested within a level 1 chapter (e.g., individual topics within “Prepared Remarks”).
- Level 3+: Further nesting is possible for finer granularity.
Only level 1 chapters are returned by default. To retrieve deeper levels, use the
levels query parameter. Multiple levels can be requested at once, e.g. levels=1,2.Example
A typical earnings call has two top-level sections, prepared remarks and a Q&A, each containing more specific segments at level 2.Level 1
levels=2 for the same audio returns the subsections within each top-level chapter.
Level 2
Interpreting timestamps
ThestartTimestamp and endTimestamp fields are in seconds from the beginning of the recording.
Notice how the level 2 chapters nest within the timestamp ranges of their level 1 parents. “CEO Opening Remarks” and “CFO Financial Review” both fall within the “Prepared Remarks” window (0–1542s), while “Analyst Questions” falls within “Q&A” (1542–3120s). You can use this relationship to build a tree structure for navigation.
