---
agentTools:
  projectIndex: https://developers.burodeingresos.com/llms.txt
---

# 📤 Async exports API and faster sandbox

<br />

* **New** — **Async exports API** — Use `POST /exports` with a `bulk_id` and selected entities to kick off a background export, receive an `export.completed` webhook when the CSV files are ready, and call `GET /exports/{export_id}` to retrieve signed download links.
* **Improved** — **Sandbox faster** — sandbox verifications now complete in under a second (previously \~12 s). No changes to the API or webhooks — only the speed.
* **Improved** — **Richer identity coverage** — profile data now includes RFC and full-name identifiers much more often for historical CURPs, and repeat verifications can be resolved with less latency.

## 📤 Async exports API

You can now export results for a completed bulk directly over the API, without going through the Console:

* `POST /exports` accepts a `bulk_id` and the entities you want to export (for now: `profile` and `invoices`) and returns an `export_id` while the export runs in the background.
* When the export finishes, a signed webhook with event `export.completed` is sent to your configured webhook URL, summarizing the generated files (name and row counts) for that `export_id`.
* `GET /exports/{export_id}` lets you poll the status and retrieve fresh signed download URLs for each CSV file whenever you need them.

Files are generated from the same data that powers `GET /profile/{identifier}` and `GET /invoices/{identifier}`, but in bulk, so large bulks can be retrieved with a single export instead of many individual reads. See the Exports section in the developer docs for full request/response examples.

## ⚡ Sandbox speed and data quality

* Sandbox traffic has been optimized so that sandbox verifications complete in under a second end-to-end, versus \~12 seconds before. The request/response shapes and webhook events are unchanged — only latency is lower.
* Profile data is now richer for the vast majority of historical CURPs: RFC and full-name information is available much more often, and new and repeat verifications for those identifiers can reuse it, so in many cases responses are faster to serve.