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

# MCP enrichment reference

> Enrich, qualify, research, and transform Landbase datasets through MCP.

Enrichment and related workflows can consume credits and may take time. Use the returned schema and quote rather than assuming fields or costs.

## For AI agents

* Call `get_dataset_fields` before choosing fields.
* Preview the dataset and show the expected quote before paid work.
* Use `confirm: true` only after explicit approval for quote-gated operations.
* Preserve `workflow_run_id` and use `get_workflow_run` until terminal status.

## Enrich a dataset

```text theme={null}
For dataset ds-7f3c2a, show the available company fields and the credit quote for industry and size range. Do not start until I approve.
```

```json theme={null}
{
  "account_id": "5093",
  "dataset_id": "ds-7f3c2a",
  "company_fields": ["industry", "size_range"],
  "confirm": true
}
```

Use `enrich_dataset` after approval. `qualify_dataset` and `research_dataset` follow the same quote-then-confirm pattern. `transform_dataset` changes the shape or filters rows; `run_workflow` starts a generic workflow. Poll all asynchronous work with `get_workflow_run`.

## Enrichment workflow

Enrichment normally operates on a matched dataset or another dataset whose
fields provide a reliable company or person anchor. First call
`get_dataset_fields`, then request the fields using the names returned for that
dataset. Do not assume that a field exists across dataset types.

The first `enrich_dataset` call should omit confirmation when the operation is
quote-gated. Show the returned row count, fields, and credit estimate. Only
repeat the same request with `confirm: true` after explicit user approval.

```json theme={null}
{
  "account_id": "5093",
  "dataset_id": "ds-7f3c2a",
  "company_fields": ["industry", "size_range"]
}
```

After approval:

```json theme={null}
{
  "account_id": "5093",
  "dataset_id": "ds-7f3c2a",
  "company_fields": ["industry", "size_range"],
  "confirm": true
}
```

Use the returned `workflow_run_id` to poll. When the run succeeds, continue
with its output dataset ID rather than assuming the source dataset was updated
in place.

## Related workflow types

* `qualify_dataset` scores rows against explicit criteria and may consume credits.
* `research_dataset` performs a quoted research task against supplied context fields.
* `transform_dataset` filters or reshapes data and is the read-only way to limit rows before paid work.
* `run_workflow` is the generic entry point when a specialized tool does not cover the operation.

## Related pages

* [MCP limits and credits](https://www.landbase.com/docs/guides/mcp/limits-and-credits.md)
* [Continue a search or workflow](https://www.landbase.com/docs/guides/mcp/continue-async-work.md)
* [MCP prompt cookbook](https://www.landbase.com/docs/guides/mcp/prompt-cookbook.md)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.