Skip to main content
GET
Get Tasks by Agent Handle

Get Tasks by Agent Handle

Fetch all tasks assigned to a specific agent/pilot using their handle. This returns full task metadata including user info, input payloads, call output, transcripts, and execution details.
Prerequisites: Make sure you have your API Token ready. See Authentication for details.
Response versions: the examples panel has two tabs. v1 is the original legacy payload. v2 is the current payload this endpoint returns — same envelope (count, status_code, message, data), but each task object carries extra fields and output.post_call_data uses schema_version: 2. Build new integrations against v2 and read post_call_data.schema_version before parsing analytics.

Path Parameters

You can get the agent_handle by hitting the Get All Agents API. The handle field in the response is your Agent Handle.

Query Parameters

count is the total number of tasks for the agent, not the size of the current page — data holds at most page_size items. Page through until you have collected count tasks.

Headers

Response Fields

Task Object Fields

Task Object Fields — v2 Additions

Present in v2 responses only. status also gains the in_progress value.
user_info and task are objects in both versions, but in v2 they come back empty ({}) for calls not created from a lead payload with an objective. Do not assume task.objective exists.

Output Object Fields

output is only fully populated once the call finishes. While status is in_progress, v2 returns a partial output with just call_type, call_status and call_end_reason.

Output Object Fields — v2 Additions

post_call_data — v2 (schema_version: 2)

In v1 post_call_data was a flat {summary, outcome, sentiment} object. In v2 it is grouped:

follow_up.policy

post_call_usage

Every numeric value under post_call_usage (call_duration_s, raw_counters, sku_deltas) is a string, not a number. Cast before doing arithmetic.

Common Error Codes

Code Examples

Best Practices

  1. Agent Handle: Use the exact agent handle from the Get All Agents response
  2. Transcript Analysis: Use the transcript field in the output for detailed conversation analysis
  3. Retry Logic: Check failure_count and last_failure_reason for failed tasks to understand issues
  4. Recording Access: Store recording_url from output for compliance and quality review
  5. Security: Keep API tokens secure and rotate them regularly

Authorizations

Authorization
string
header
required

Format: Token

Headers

Org-Handle
string
required

Organization domain handle

Example:

"unpod.tv"

Path Parameters

agent_handle
string
required
Example:

"space-agent-8qmk42nslp91wrh3dz7btxc4"

Query Parameters

page
integer
Example:

1

page_size
integer
Example:

20

Response

List of tasks for the agent