# Get the analysis

`GET https://api.eesi.ai/v1/speech/transcriptions/{session_id}/intelligence`

The analysis of a recording: how each line was delivered, the
audience's reactions, each speaker's profile and arc, the moments worth a look,
and a written summary.

`status` is `queued`, `running`, `complete`, `failed`, `unavailable` (the
recording has no audio or no speech) or `none` (never analyzed). While it runs,
`stage` and `progress` say how far it is, and the document fills in as it goes:
measurements first, then what each line sounds like, then the summary. Poll every
few seconds until `complete`.

Every signal says what backs it in `source`: `measured` (read from the audio
signal), `heard` (from listening to the line), or `both`.

Parameters:

- `session_id` (path, required)

Responses: 200, 404, 422.

Schemas: https://docs.eesi.ai/api-reference/public.json