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

# Dashboard

> A tour of the web app: what client teams see, and the operational surfaces that run the machine.

The dashboard is two products sharing one shell: a **consumption surface** for client teams (chat, visualisation, upload) and an **operations surface** for the people running ingestion and extraction. What you see is determined by your [role](/platform/multi-tenancy).

## For client teams

| Surface             | What it's for                                                                                                                                |
| ------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- |
| **Chat**            | The default landing surface: conversational access to extracted knowledge with citations and persistent threads. See [Chat](/platform/chat). |
| **Visualize**       | Process and entity visualisation over the extracted knowledge.                                                                               |
| **Document upload** | Drop files or ZIP archives into the [Document Inbox](/platform/documents#document-inbox); any member can feed the corpus.                    |
| **Settings**        | Org configuration; admins additionally manage members and API keys.                                                                          |

## For operators

The operational surfaces expose every layer of the [architecture](/platform/architecture), in the same order data flows through it:

| Surface                           | Layer     | What you do there                                                                                                                     |
| --------------------------------- | --------- | ------------------------------------------------------------------------------------------------------------------------------------- |
| **Connectors**                    | Data in   | Create and configure sources, trigger syncs, watch per-run counters and error reasons.                                                |
| **Document monitor**              | Data in   | Per-file processing status, skip reasons, reprocess controls.                                                                         |
| **Raw records / Events / Actors** | Data      | Inspect exactly what was ingested and how it normalized.                                                                              |
| **Extraction runs**               | Analysis  | Trigger and monitor runs; drill into per-stage outputs.                                                                               |
| **Inference logs**                | Analysis  | The actual prompts, responses, and token counts of every LLM call.                                                                    |
| **Libraries / Rules**             | Knowledge | Browse versions, review rules (confirm / modify / reject).                                                                            |
| **Query Explorer**                | Querying  | Run queries with [explain mode](/platform/query#explain-mode): per-result signal attribution for debugging "why did it answer that?". |

<Note>
  The operational surfaces are intentionally transparent down to the prompt level. When a client asks "where did this rule come from?", an operator can walk the answer live: rule → inference log → the exact prompt and the events it saw. The dashboard is the [provenance chain](/platform/provenance) made clickable.
</Note>

## Design notes

* **Live where it matters.** Sync runs and extraction runs poll while in progress and relax when idle; you watch progress without refreshing.
* **Role checks mirror the backend.** Navigation hides what you can't access, but every page is also enforced server-side; the UI gating is convenience, not security.
* **Export over lock-in.** Rules export to CSV, and the API exposes everything the dashboard shows.

## Related

<CardGroup cols={2}>
  <Card title="Multi-tenancy & Roles" icon="building" href="/platform/multi-tenancy">
    Exactly which role sees which surface.
  </Card>

  <Card title="Architecture" icon="layer-group" href="/platform/architecture">
    The layers these surfaces operate.
  </Card>
</CardGroup>