> ## 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.

# v2 conventions

> Rules that apply across every AgentSource v2 endpoint: path versioning, single endpoints instead of bulk routes, and the sync/async pairing.

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

These rules hold across the whole v2 surface, so they are not repeated on individual endpoint pages.

## Path versioning

Every endpoint moves under the `/v2` path, replacing its `/v1` equivalent.

```
POST /v1/businesses/firmographics/bulk_enrich   →   POST /v2/businesses/firmographics/enrich
```

## No more bulk endpoints

A single endpoint accepts either one ID (a string) or a list of IDs. The separate `/bulk_enrich` routes are removed.

<CodeGroup>
  ```json Single ID theme={null}
  { "business_id": "a34bacf839b923770b2c360eefa26748" }
  ```

  ```json List of IDs theme={null}
  { "business_id": ["a34bacf839b923770b2c360eefa26748", "8adce3ca1cef0c986b22310e369a0793"] }
  ```
</CodeGroup>

<Note>
  Request and response schemas are still being finalized. The example above shows the calling pattern, not the confirmed field names.
</Note>

## Sync vs. async pairing

Each enrichment exposes two endpoints:

| Variant   | Path         | Input                                | Use when                         |
| :-------- | :----------- | :----------------------------------- | :------------------------------- |
| **Sync**  | `.../enrich` | A single ID or a list of up to 50    | You want results in the response |
| **Async** | `.../job`    | A `list_id` from an uploaded dataset | Large runs, up to 10,000 records |

<Warning>
  The "single ID or a list of IDs" rule above applies to **sync** endpoints. Async `.../job` endpoints take a `list_id` from the [dataset upload endpoint](/v2/jobs/upload_entity_id_dataset) and do not accept inline IDs — see the [end-to-end guide](/v2/async-guide).
</Warning>

Both are `POST`. The async variant is subject to the async job limits: **max 10K input records** and a **24-hour maximum run time**.

## Related

<Columns cols={2}>
  <Card title="Async jobs" icon="clock" href="/v2/async-jobs">
    How job submission, polling, and retrieval work
  </Card>

  <Card title="Migrating from v1" icon="arrow-right-arrow-left" href="/v2/migration-from-v1">
    Apply these conventions to your existing calls
  </Card>
</Columns>