Skip to main content

Recorded speech features

Recorded calls pass through audio preparation, transcription, and the post-processing features enabled for the account.

Audio input and preparation​

Speechlytics accepts audio through base64 JSON input, a downloadable URL, scheduled provider integrations, FTP/SFTP workflows, and the live-call finalizer. The configured preprocessing pipeline can:

  • validate the media format and minimum duration;
  • split or preserve stereo channels;
  • remove an initial welcome message before transcription;
  • detect silence, noise, music, and laughter;
  • encrypt stored payloads when deployment encryption is enabled.

The request body limit is approximately 300 MB. URL submission avoids expanding binary audio into base64 and is preferable for large files.

Language handling​

Set language on the process request when the spoken language is known. Use 0 (Auto) when it is not. Account defaults can replace Auto, so confirm the configured account language if automatic detection does not behave as expected.

The API advertises more than 100 language values in the Language schema. Actual recognition quality depends on the deployed speech model.

Mono and stereo​

Stereo audio is separated into left and right transcription channels. Mono audio uses the shared Both channel. Speaker-specific analytics are only reliable when the source preserves speakers on separate channels.

Keyword matching and scoring​

Configured keywords and phrases are matched after transcription. A keyword can target a channel, carry a score, be exclusive, mark a call out of scope, and contain reporting tags. Results include matched keywords, tag frequencies, and the aggregate keyword score.

Read Understanding call scores for the product view and Scoring and analytics for calculation details.

AI analytics​

Depending on account configuration, post-processing can add:

  • conversation summaries for Sales or Customer Support workflows;
  • topics, sentiment, emotions, intent, business value, and performance signals;
  • named entities and structured call facts;
  • translated transcript segments;
  • custom QA questions with pass/fail, evidence, detected names, and extra fields;
  • sensitive-content masking before downstream matching and analytics.

These fields are returned by call status and call list endpoints when present.

Agent intelligence​

When call metadata identifies an agent and Intelligence is enabled, Speechlytics aggregates calls into agent scorecards, performance trends, sentiment and topic distributions, and call history. See the Intelligence module.

Results and filtering​

Use:

  • GET /api/v1/calls/{id}/status for one call;
  • GET /api/v1/calls for paginated calls filtered by date, filename, topic, sentiment, tag, content, or cluster;
  • GET /api/v1/analytics for paginated analytics;
  • GET /api/v1/topics and /api/v1/sentiments for account-level distributions;
  • GET /api/v1/intelligence for agent-level analytics.

See the API reference for exact query parameters.