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

# Data Platform Overview

> Knowledge stores, memory, connectors, and context assembly for your agents.

The Data Platform solves two problems for agents: giving them knowledge to search (product docs, policies, internal wikis) and memory to retain (conversation history, user preferences, past decisions). Without these, every agent invocation starts from zero.

It provides searchable **knowledge stores**, scoped **memory**, **connectors** that keep content in sync with external sources, and **context bundles** that assemble grounded context for a run.

## Capabilities

<CardGroup cols={2}>
  <Card title="Knowledge Stores" icon="book" href="/docs/data-platform/knowledge-stores">
    Ingest content and search it semantically.
  </Card>

  <Card title="Memory" icon="brain" href="/docs/data-platform/memory">
    Record and recall agent memory across scopes.
  </Card>

  <Card title="Connectors" icon="plug" href="/docs/data-platform/connectors">
    Sync data from Amazon S3 into knowledge stores.
  </Card>

  <Card title="Documents & Uploads" icon="file-arrow-up" href="/docs/data-platform/documents">
    Upload files for automatic chunking and embedding.
  </Card>

  <Card title="Syncs" icon="arrows-rotate" href="/docs/data-platform/syncs">
    Monitor ingestion runs across your environment.
  </Card>

  <Card title="Context Bundles" icon="box" href="/docs/data-platform/context-bundles">
    Assemble grounded context for a task run.
  </Card>
</CardGroup>

## Base URL

All Data Platform requests use the REST API:

```
https://api.hexelstudio.com/data/v1/...
```

Authenticate with a Bearer token. See [Authentication](/docs/getting-started/authentication).

<Note>
  The Data Platform is accessed via the REST API and Console only. The Python and TypeScript SDKs cover Compute, Orchestration, and Tools; Data Platform examples in these docs use `curl` against the REST API.
</Note>

## Errors

Data Platform endpoints return a consistent JSON error body: `{ "code": "<error_code>", "message": "<human_message>" }`. Permission errors include the required permission in `details`.

| `code`                   | HTTP | When                                                                  |
| ------------------------ | ---- | --------------------------------------------------------------------- |
| `AUTH.REQUIRED`          | 401  | Missing or invalid credentials.                                       |
| `AUTH.PERMISSION_DENIED` | 403  | Authenticated but lacking the required `data-platform:*` permission.  |
| `RESOURCE.NOT_FOUND`     | 404  | The store, document, datasource, or memory item doesn't exist.        |
| `VALIDATION`             | 400  | Invalid parameters (out-of-range chunk size, bad query length, etc.). |
| —                        | 429  | Rate limit exceeded (10,000 requests/hour per organization).          |

## Security

| Concern        | How it's handled                                                                                                           |
| -------------- | -------------------------------------------------------------------------------------------------------------------------- |
| Authentication | Every request uses a short-lived Bearer token.                                                                             |
| Authorization  | Each operation requires a `data-platform:<resource>:<action>` permission (for example, `data-platform:knowledge:write`).   |
| Scoping        | Knowledge stores, memory, datasources, and documents are scoped to your organization, workspace, and environment.          |
| Storage        | Content and embeddings are stored on platform-managed infrastructure; you don't provision or manage the underlying stores. |

## Rate limits

| Scope            | Limit                  |
| ---------------- | ---------------------- |
| Per organization | 10,000 requests / hour |

## Pagination

List endpoints accept `page` (starting at 1) and `page_size` (1–100, default 20).

```bash theme={"dark"}
curl "https://api.hexelstudio.com/data/v1/knowledge?page=1&page_size=20" \
  -H "Authorization: Bearer $TOKEN"
```

## Next steps

Start with [Knowledge Stores](/docs/data-platform/knowledge-stores) to ingest and search your first content.