Skip to main content

Calls and call details

The Calls area is the primary place to find processed conversations, monitor work in progress, and inspect transcript and analytics results.

Calls list​

Each row summarizes a call with identifying metadata, time, duration, processing status, scores, and available analytics. The exact columns depend on the account's capabilities and the data present on the call.

Common filters include:

  • date range;
  • filename;
  • topic and sentiment;
  • keyword tag;
  • transcript content;
  • processing cluster where applicable.

The REST equivalent is GET /api/v1/calls, which returns calls visible to the authenticated account and accepts the same major filter dimensions.

Processing states​

StateMeaning
QueuedThe call is waiting to be processed.
InProgressAudio or transcript processing is active.
ProcessedTranscription and enabled post-processing are complete.
FailedProcessing stopped with an error.
QuotaLimitThe account did not have enough remaining minutes.
RedactionStarted / RedactionEndedSensitive-content protection is active or complete.

Use GET /api/v1/calls/{id}/status to retrieve the latest state and result for a specific call.

Call detail layout​

A completed call can contain:

Audio and transcript​

  • audio playback and duration;
  • left, right, or shared transcript channels;
  • translated segments when available;
  • silence, music, noise, and laughter measurements;
  • speaker talk-share and frequent-word summaries.

Keywords and scores​

  • matched keyword definitions and channel matches;
  • tag frequencies;
  • rule-based keyword score;
  • AI-derived analytics score;
  • scope or exclusion indicators.

Conversation analytics​

  • summary and structured call analysis;
  • topics, sentiment, emotions, and intent;
  • strengths, risks, opportunities, and next steps appropriate to the account's analysis profile;
  • Custom Question verdicts and evidence;
  • agent attribution and external provider identifiers.

Result sections are optional. A missing section normally means the capability was unavailable, the source audio did not support it, or the stage did not produce a result.

Mono and stereo results​

Stereo recordings can preserve separate agent and customer channels. Mono recordings use a shared channel, so Speechlytics does not present unmeasured per-speaker metrics as if they were separated.

Access and visibility​

Call endpoints are tenant-scoped. A valid token cannot retrieve another account's call by guessing its numeric ID. Calls outside the token's visibility are returned as not found.

Retrieve results programmatically​

curl "https://api.example.com/api/v1/calls/12345/status" \
  --header "Authorization: Bearer $SPEECHLYTICS_TOKEN"

The status result is the most complete single-call API representation. See the API reference for all fields and the feature guide for optional processing stages.