Dasha Quick Start GuidePlatform REST API

Job analysis API

Use the job analysis API to analyze a job's recording and transcript. The endpoint supports the built-in quality analysis and custom analysis with a prompt and JSON Schema.

The API stores each analysis request and its result. By default, repeated requests reuse an existing queued, processing, or completed analysis for the same job, prompt, and schema. Use force=true to create a new analysis.

Authentication

Use regular Dasha customer authentication:

Authorization: Bearer <DASHA_API_KEY>

The token's customer must own the job. The account must also have job analysis enabled.

Start analysis

POST /api/v1/jobs/{jobId}/analyze

Query parameters:

Parameter	Type	Default	Description
`force`	boolean	`false`	Create a new analysis instead of reusing an existing queued, processing, or completed analysis for the same inputs.
`sync`	boolean	`false`	Keep the HTTP request open while the service polls for a terminal result.
`syncTimeoutSeconds`	integer	`60`	Maximum wait time when `sync=true`. Accepted range: `1` to `120`.

The request body is optional. Omit it, or send {}, to run the built-in quality analysis:

curl -X POST "$DASHA_API_BASE_URL/api/v1/jobs/$JOB_ID/analyze" \
  -H "Authorization: Bearer $DASHA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

Send both prompt and schema to run a custom analysis:

curl -X POST "$DASHA_API_BASE_URL/api/v1/jobs/$JOB_ID/analyze?sync=true&syncTimeoutSeconds=90" \
  -H "Authorization: Bearer $DASHA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Analyze the conversation. Return an NPS score from 0 to 10. Also report whether the agent greeted the customer and asked if it was a good time to talk.",
    "schema": {
      "type": "object",
      "properties": {
        "nps": {
          "type": "integer",
          "minimum": 0,
          "maximum": 10
        },
        "scorecard": {
          "type": "object",
          "properties": {
            "greetedCustomer": { "type": "boolean" },
            "askedGoodTimeToTalk": { "type": "boolean" }
          },
          "required": ["greetedCustomer", "askedGoodTimeToTalk"]
        }
      },
      "required": ["nps", "scorecard"]
    }
  }'

Custom analysis requests must meet these rules:

Provide prompt and schema together.
Keep prompt at 8000 characters or fewer.
Keep serialized schema at 32000 characters or fewer.
Use object as the root schema type.
Use only these schema keywords: type, properties, required, items, enum, description, additionalProperties, minItems, maxItems, minLength, maxLength, minimum, and maximum.

Response

POST returns 202 Accepted while analysis is queued or processing. It returns 200 OK when it can return a terminal result, including a cached completed result or a sync request that finishes before the timeout.

{
  "analysisId": "0b716598-3d3a-4fd9-8b8c-4a6b27710f3a",
  "jobId": "019df6cf-6526-7277-b742-cb19ba6fdfec",
  "analysisType": "custom",
  "status": "completed",
  "createdAt": "2026-05-25T08:10:11.123456+00:00",
  "updatedAt": "2026-05-25T08:10:35.123456+00:00",
  "result": {
    "nps": 7,
    "scorecard": {
      "greetedCustomer": true,
      "askedGoodTimeToTalk": false
    }
  },
  "error": null
}

Response fields:

Field	Type	Description
`analysisId`	string	Unique analysis request ID.
`jobId`	string	Job ID from the URL.
`analysisType`	string	`quality` for built-in analysis or `custom` for prompt/schema analysis.
`status`	string	`queued`, `processing`, `completed`, or `failed`.
`createdAt`	string	Analysis creation timestamp.
`updatedAt`	string	Last analysis update timestamp.
`result`	object or null	JSON result when `status` is `completed`; otherwise `null`.
`error`	object or null	Error details when `status` is `failed`; otherwise `null`.

Error response codes:

Status	Meaning
`400 Bad Request`	The request body, schema, or sync timeout is invalid.
`403 Forbidden`	The token is not authorized, or the account cannot use job analysis.
`404 Not Found`	The job or analysis does not exist for the authenticated customer.

Retrieve analyses

List all stored analyses for a job:

curl "$DASHA_API_BASE_URL/api/v1/jobs/$JOB_ID/analyze" \
  -H "Authorization: Bearer $DASHA_API_KEY"

Response:

{
  "items": [
    {
      "analysisId": "0b716598-3d3a-4fd9-8b8c-4a6b27710f3a",
      "jobId": "019df6cf-6526-7277-b742-cb19ba6fdfec",
      "analysisType": "custom",
      "status": "completed",
      "createdAt": "2026-05-25T08:10:11.123456+00:00",
      "updatedAt": "2026-05-25T08:10:35.123456+00:00",
      "result": {
        "nps": 7,
        "scorecard": {
          "greetedCustomer": true,
          "askedGoodTimeToTalk": false
        }
      },
      "error": null
    }
  ]
}

Get one analysis by ID:

curl "$DASHA_API_BASE_URL/api/v1/jobs/$JOB_ID/analyze/$ANALYSIS_ID" \
  -H "Authorization: Bearer $DASHA_API_KEY"

Delete all stored analyses for a job:

curl -X DELETE "$DASHA_API_BASE_URL/api/v1/jobs/$JOB_ID/analyze" \
  -H "Authorization: Bearer $DASHA_API_KEY"

DELETE returns 204 No Content.

Found a mistake? Let us know.

Job analysis API

Job analysis API

Authentication

Start analysis

Response

Retrieve analyses

Enroll in beta