> For the complete documentation index, see [llms.txt](https://docs.seemoredata.io/external-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.seemoredata.io/external-docs/fundamentals/settings/api-access/tool-reference.md).

# MCP Tool Reference

## Overview

When you connect an AI assistant over MCP, it gains access to the tools below. You don't call these directly — you ask questions in natural language and your assistant chooses the right tools. This reference helps you understand what your assistant can do and write more effective prompts.

{% hint style="info" %}
**Required role:** [Viewer](/external-docs/fundamentals/settings/user-roles.md) or higher for every read tool. Creating, updating, or deleting custom tags requires [Editor](/external-docs/fundamentals/settings/user-roles.md) or higher.
{% endhint %}

Throughout this reference, a **compute unit** is a Snowflake **warehouse**, and an **account** is one of your connected warehouse or BI integrations. Every tool returns only the data your role and integration access already permit.

***

## Account & discovery

| Tool                  | What it does                                                                                                                                                                                    |
| --------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `get-account-summary` | Returns your account metadata and the integrations you can see. A good first call to orient your assistant on what's in scope.                                                                  |
| `search-assets`       | Full-text search across all data assets — tables, jobs, dashboards, queries, and terms — with filters for entity type, integration, and database. Returns matches with aggregated facet counts. |
| `get-asset`           | Fetches a single asset by id, with rich detail (for tables: columns, row count, size, and cost; for jobs and queries: the definition and schedule).                                             |
| `get-lineage`         | Walks the upstream and/or downstream lineage graph from an asset, with optional cost, usage, and freshness metrics per node over a chosen time window.                                          |

***

## Cost & spend trends

Each tool returns a time-series of cost broken into buckets, with grouping and filtering that mirror the Seemore UI. Windows can be a preset range or an explicit start/end date (up to 31 days).

| Tool                               | What it does                                                                                   |
| ---------------------------------- | ---------------------------------------------------------------------------------------------- |
| `get-compute-units-cost-overtime`  | Warehouse (compute) cost over time, grouped by account, domain, warehouse, or tag.             |
| `get-workloads-cost-overtime`      | Snowflake query workload cost over time, grouped by warehouse, user, or query-tag value.       |
| `get-storage-cost-overtime`        | Storage cost over time, grouped by account, database, or tag.                                  |
| `get-cloud-services-cost-overtime` | Per-cloud-service cost over time (e.g. Auto-Clustering, Materialized Views, Snowpipe, Cortex). |

***

## Warehouses (compute units)

| Tool                               | What it does                                                                                                                                                                                                                                         |
| ---------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `list-compute-units`               | Lists your warehouses with their configuration — size, auto-suspend, scaling policy, and resource constraints.                                                                                                                                       |
| `list-compute-unit-events`         | Shows the configuration change history for a warehouse (size, auto-suspend, scaling-policy changes), including whether each change came from a user or a Seemore automation (Smart Pulse, Burst Protector, Auto Shutdown, Auto Scaler, Live Tuning). |
| `get-compute-unit-events-metadata` | Lists the change types and change sources available for filtering a warehouse's event history.                                                                                                                                                       |

***

## Budgets

| Tool                   | What it does                                                                             |
| ---------------------- | ---------------------------------------------------------------------------------------- |
| `list-budgets`         | Lists your configured budgets with current spend, average daily cost, and percent spent. |
| `get-budget`           | Returns a single budget with a freshly computed spend snapshot.                          |
| `get-budget-drilldown` | Returns a budget's day-by-day spend and per-resource cost attribution over its period.   |

***

## Anomalies & insights

| Tool              | What it does                                                                                                                                                                   |
| ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `list-anomalies`  | Lists detected cost anomalies (unexpected spikes or drops), with expected vs. actual cost and severity. Filter by status and severity.                                         |
| `analyze-anomaly` | **AI agent.** Runs Seemore's Snowflake Cortex–powered root-cause analysis on a single anomaly and streams its reasoning back. Supports follow-up questions in the same thread. |
| `list-insights`   | Lists actionable waste-reduction and optimization insights (unused tables, orphan tables, bloated storage, inactive users, and more) with projected savings.                   |

***

## Queries & optimization

| Tool                | What it does                                                                                                                                                                                                                    |
| ------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `get-query-history` | Filterable, paginated search over warehouse query history with rich per-query metadata (cost, duration, spillage, destination table, and more). Requires a time window of up to 7 days, which may target any historical period. |
| `similar-queries`   | Finds other runs of the same query (matched by statement text) so you can spot cost or runtime regressions across executions.                                                                                                   |
| `analyze-query`     | **AI agent.** Runs Seemore's Snowflake Cortex–powered query-optimization analysis on a single query and streams its reasoning back. Supports follow-up questions in the same thread.                                            |

{% hint style="info" %}
`analyze-anomaly` and `analyze-query` invoke Seemore's AI agents and stream their analysis as it's produced. Your assistant relays the reasoning and final recommendation back to you in the conversation.
{% endhint %}

***

## Tags & tag rules

These tools cover both Snowflake-native object tags and Seemore custom tags. For background on how tags are extracted and used, see [Query Tags](/external-docs/fundamentals/our-features/observability/query-tags.md) and [Tags Management](/external-docs/fundamentals/settings/tags-management.md).

{% hint style="info" %}
**Required role:** the three write tools below (`create-custom-tag`, `update-custom-tag`, `delete-custom-tag`) require [Editor](/external-docs/fundamentals/settings/user-roles.md) or higher. All other tag tools are read-only and available to [Viewer](/external-docs/fundamentals/settings/user-roles.md) or higher.
{% endhint %}

| Tool                                     | Type      | What it does                                                                                                                           |
| ---------------------------------------- | --------- | -------------------------------------------------------------------------------------------------------------------------------------- |
| `find-tag`                               | Read      | Searches for a tag by name across every surface (query, object, and custom) in one call. The best starting point for any tag question. |
| `list-custom-tags`                       | Read      | Lists custom-tag definitions, including any value transformer and fallback value.                                                      |
| `list-query-tag-names`                   | Read      | Lists query-tag keys and their observed values — useful for filtering query and workload cost data.                                    |
| `list-object-tag-names`                  | Read      | Lists object-tag keys (on databases, schemas, tables, warehouses, users, and accounts) and their observed values.                      |
| `list-custom-query-tag-rules`            | Read      | Lists the rules that extract a custom tag's value from query-history columns.                                                          |
| `list-custom-object-tag-rules`           | Read      | Lists the rules that assign a custom tag to assets matching a set of conditions.                                                       |
| `list-custom-object-tag-rule-predicates` | Read      | Lists the individual conditions (asset type, column, operator, values) that make up an object-tag rule.                                |
| `create-custom-tag`                      | **Write** | Creates a custom-tag configuration, with an optional value transformer and fallback value.                                             |
| `update-custom-tag`                      | **Write** | Updates an existing custom-tag configuration in place.                                                                                 |
| `delete-custom-tag`                      | **Write** | Deletes a custom-tag configuration and all of its attached rules. This is destructive and cannot be undone.                            |

{% hint style="warning" %}
`delete-custom-tag` permanently removes a tag configuration and every rule attached to it. A well-behaved AI client will ask you to confirm before running it — review the request before approving.
{% endhint %}