---
title: "Analyze Single File"
method: POST
path: "/api/v1/offline/processing"
tags: ["Processing"]
---

# Analyze Single File

`POST /api/v1/offline/processing`

For the best results, use audio files with a sample rate of 16 kHz and 16 bits per sample.

## Request body

- Pipeline — Main configuration for audio processing
  - `config` RecognitionConfig, required — Configuration for audio processing pipeline
    - `language_code` 'en-US' | 'he-IL' | 'hi-IN' | 'es-ES' | 'fr-FR' | 'de-DE' | 'it-IT' — Primary language of the audio file
    - `transcript_model` 'openai_whisper' — Speech-to-text model to use for transcription
    - `emotions_model` string — Model to use for emotion detection. Set to null to disable
    - `emotions_alignment` boolean — Enable temporal alignment of emotions with speech
    - `emotions_diarization` boolean — Enable speaker-based emotion diarization
    - `channels` 1 | 2 — Number of audio channels (1 for mono, 2 for stereo)
    - `insights` object — Configure which insights to generate and their options
      - `agent_actions` object
        - `prompt_customization` string — Custom prompt for agent actions analysis
      - `csat` object
        - `prompt_customization` string — Custom prompt for CSAT analysis
      - `customer_requests` object
        - `prompt_customization` string — Custom prompt for customer requests analysis
      - `division` object
        - `prompt_customization` string — Custom prompt for conversation division analysis
      - `keywords` object
        - `prompt_customization` string — Custom prompt for keywords extraction
      - `label` object
        - `prompt_customization` string — Custom prompt for labeling
        - `labels` string[] — List of labels to classify the conversation
      - `problems` object
        - `prompt_customization` string — Custom prompt for problems detection
      - `questions` object
        - `prompt_customization` string — Custom prompt for questions analysis
      - `sentiment` object
        - `prompt_customization` string — Custom prompt for sentiment analysis
      - `summary` object
        - `prompt_customization` string — Custom prompt for summary generation
      - `topics` object
        - `prompt_customization` string — Custom prompt for topics extraction
    - `webhook` string, uri, nullable — URL to receive processing results when complete
  - `audio` RecognitionAudio, required — Audio file information
    - `uri` string, uri, required — Publicly accessible URL to the audio file
  - `call_data` RecognitionMeta — Optional metadata about the call
    - `id` string — Your internal identifier for this call
    - `timestamp` string, date-time — When the call occurred
    - `speakers` object[] — Information about speakers in the call
      - `id` string, required — Speaker identifier
      - `label` string — Display name for the speaker
      - `role` 'agent' | 'customer', required — Role of the speaker

## Response `200`

Successful Response

- ResultId
  - `result_id` string — Identifier to retrieve the analysis results

## Other responses

- `422` — Validation Error

## Changes

> 7 revisions in range; 1 could not be searched.

- **2025-01-13** `5b83f89b67ac` — 6 breaking, 3 warning, 7 info
  - removed the enum value `agent-actions` of the request property `config/insights/items/`
  - removed the enum value `customer-requests` of the request property `config/insights/items/`
  - removed the enum value `descriptor-default` of the request property `config/transcript_model`
  - removed the enum value `descriptor-hebrew` of the request property `config/transcript_model`
  - …12 more
- **2024-10-31** `7dacbc6fdf9a` — 1 breaking, 1 info
  - the `config/insights` request property type/format changed from `object`/`` to `array`/``
  - the `insights` request property default value changed from `{}` to ``
- **2024-10-31** `19c8adc9d2aa` — 3 info
  - the `transcript_model` request property const value `descriptor-hebrew` was removed
  - the `transcript_model` request property default value changed from `descriptor-hebrew` to `descriptor-default`
  - added the new `descriptor-default` enum value to the request property `config/transcript_model`
- …earlier changes not shown

[Full history](https://skmtc.dev/descriptor-ai/apis/m-path/changes/api/v1/offline/processing/post.md)

---

[API](https://skmtc.dev/descriptor-ai/apis/m-path.md) · [All operations](https://skmtc.dev/descriptor-ai/apis/m-path/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/descriptor-ai/m-path/revisions/cb91b959a95f/schema)