> ## Documentation Index
> Fetch the complete documentation index at: https://developers.explorium.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Async job infrastructure

> How AgentSource v2 async endpoints work: job submission, status polling, cancellation, dataset upload, and the input and runtime limits.

<Info>
  **Beta.** AgentSource v2 is in beta. Details may still change before GA.
</Info>

## Overview

Every `.../job` endpoint in v2 runs through a shared async job layer. Instead of returning results inline, a job endpoint accepts your input and returns a job reference. You then poll for state and collect results.

<Steps>
  <Step title="Submit">
    Upload a CSV of entity IDs, then call any `.../job` endpoint with the returned `list_id` — for example [Firmographics (async)](/v2/businesses/enrichments/firmographics_job).
  </Step>

  <Step title="Poll">
    Call [Get job status](/v2/jobs/get_job_status) with the returned job ID to check state and retrieve results once complete.
  </Step>

  <Step title="Cancel (optional)">
    Call [Cancel job](/v2/jobs/cancel_job) to stop a job that is queued or running.
  </Step>
</Steps>

## Endpoints

| Purpose                                  | Endpoint                                   |
| :--------------------------------------- | :----------------------------------------- |
| Job status and results                   | `GET /v2/jobs/status/{job_id}`             |
| List your jobs                           | `GET /v2/jobs`                             |
| Cancel a job                             | `POST /v2/jobs/cancel/{job_id}`            |
| Upload an entity ID dataset as job input | `POST /v2/async/entity-id-datasets/upload` |

Async jobs take their input **only** as an uploaded dataset — [upload your IDs](/v2/jobs/upload_entity_id_dataset) and pass the returned `list_id`. Inline IDs are not accepted. The [end-to-end guide](/v2/async-guide) walks through the whole flow.

## Limits

| Limit                                   | Value                                                                 |
| :-------------------------------------- | :-------------------------------------------------------------------- |
| Maximum input size per async enrichment | 10,000 records                                                        |
| Maximum job run time                    | 24 hours                                                              |
| Download link lifetime                  | \~1 hour — [re-poll status](/v2/jobs/get_job_status) for a fresh link |
| Result file retention                   | 7 days after the job finishes                                         |

## Which endpoints are async

Every enrichment has an async variant — see the **Async** groups under Business enrichments and Prospect enrichments in the sidebar, or the [business](/v2/businesses/enrichments/overview) and [prospect](/v2/prospects/enrichments/overview) enrichment overviews.

These are **sync only**:

* Match, fetch, and stats for [businesses](/v2/businesses/fetch_businesses) and [prospects](/v2/prospects/fetch_prospects)
* [Autocomplete](/v2/autocomplete)
* [Credits](/v2/credits/get_credits)