Skip to main content

Oncrawl MCP Server – Tools reference

Discover the Oncrawl MCP tools… access data, run analyses, and manage SEO operations directly from your AI assistant.

Updated today

The Oncrawl MCP Server exposes a set of tools that map to Oncrawl capabilities.

These tools allow AI clients to retrieve data, run analyses, and perform operations.

Context and account discovery

get_context_user

Returns the current user context, including:

  • workspaces

  • the last active project for each workspace

  • the last active crawl for each project

  • the crawl configuration for the last active crawl

This is useful when a user wants a quick overview of their Oncrawl environment before querying deeper.

Workspace and project management

workspace_operations

Supports:

  • list

  • get

Use this tool to retrieve workspace information.

project_operations

Supports:

  • list

  • get

  • create

  • delete

Use this tool to manage Oncrawl projects.

Project creation requires:

  • name

  • start_url

A project does not automatically create a crawl configuration or launch a crawl. To start crawling, you must create a crawl configuration and then run a crawl.

Crawl configuration and crawl management

crawl_config_operations

Supports:

  • list

  • get

  • create

  • update

  • delete

Use this tool to manage crawl configurations for a project.

A crawl configuration requires these mandatory fields:

  • name

  • start_url

  • user_agent

  • max_speed

  • max_depth

  • max_url

crawl_operations

Supports:

  • list

  • get

  • run

  • stop

  • progress

  • get_commands

  • command

Use this tool to list crawls, start a crawl from a project and config, stop a crawl, inspect progress, or retrieve available crawl commands.

To run a crawl, you need:

  • project_id

  • config_id

If a project does not yet have a crawl configuration, you must create one first.

scheduled_crawl_operations

Supports:

  • list

  • get

  • create

  • update

  • delete

Use this tool to manage recurring crawls for a project.

A scheduled crawl requires:

  • config_id

  • start_date

  • recurrence

Optional:

  • allow_concurrent_crawls

Crawl-over-crawl management

crawl_over_crawl_operations

Supports:

  • list

  • get

  • create

  • delete

Use this tool to manage crawl-over-crawl comparisons.

Creating a comparison requires:

  • project_id

  • crawl1

  • crawl2

Optional configuration supports URL transformation rules, such as prefix replacement.

Crawl, log, and comparison data retrieval

The MCP server distinguishes between search tools for retrieving rows and aggregate tools for grouped statistics.

Search tools

get_data_search_crawl

Searches crawl data for:

  • pages

  • links

  • clusters

  • structured_data

Use this when you want raw matching results rather than summaries.

get_data_search_coc

Searches crawl-over-crawl data.

Use this when you want matching URLs or pages from a specific comparison.

get_data_search_logs

Searches project log data for:

  • pages

  • events

When querying pages, a granularity is required:

  • days

  • weeks

  • months

Use this when you need raw log rows rather than aggregated metrics.

Aggregation tools

get_data_aggregate_crawl

Runs aggregations on crawl datasets.

Supported crawl data types:

  • pages

  • links

  • clusters

  • structured_data

Use this for grouped counts, averages, sums, cardinality, and similar statistics.

get_data_aggregate_coc

Runs aggregations on crawl-over-crawl page data.

Use this for questions such as:

  • how many pages are new, removed, or matching

  • how many pages changed status code

  • average delta metrics across matching pages

get_data_aggregate_logs

Runs aggregations on log datasets.

Supported log data types:

  • pages

  • events

For page-level logs, granularity is required:

  • days

  • weeks

  • months

Use this for grouped counts, summaries, and time-based log analysis.

Field discovery and OQL support

get_fields_crawl

Returns available fields for a crawl data type.

get_fields_coc

Returns available fields for a crawl-over-crawl dataset.

get_fields_logs

Returns available fields for log data.

These tools are especially useful before writing advanced filters or selecting

response fields.

get_oql_documentation

Returns documentation for OQL (Oncrawl Query Language).

Supported topics include:

  • syntax

  • examples

  • field filters

  • logical operators

  • data types

  • data-type-specific examples

Use this tool when building or explaining OQL filters.

URL inspection

url_details_operations

Supports:

  • get_url_details

  • get_url_headers

  • get_url_body

Use this tool to inspect a URL from a crawl, including:

  • fetch date

  • response headers

  • response body

  • status code

This is useful for troubleshooting page-level crawl behavior.

Exports, secrets, and related operations

data_export_operations

Supports:

  • list

  • get

  • create

Use this tool to manage workspace data exports.

Supported export targets include:

  • S3

  • GCS

Supported output formats include:

  • JSON

  • CSV

  • Parquet

secret_operations

Supports:

  • list

  • create

  • delete

Use this tool to manage workspace secrets used by export targets.

Supported secret types include:

  • gcs_credentials

  • s3_credentials

Ranking performance

ranking_performance_operations

Supports:

  • fields

  • aggregations

Use this tool to retrieve available ranking fields or run aggregate queries on ranking performance data for a project.

Search vs. aggregation tools

When working with Oncrawl data through MCP, it helps to choose the right type of tool.

Use a search tool when you want:

  • a list of matching pages

  • a list of links

  • raw crawl-over-crawl results

  • raw log rows

Use an aggregation tool when you want:

  • counts by status code

  • average load time

  • grouped summaries

  • distributions by template, URL type, or time period

Limits and query behavior

Some data search tools return paginated results and enforce limits.

For these tools:

  • get_data_search_crawl

  • get_data_search_coc

  • get_data_search_logs

the maximum result limit is 20 per request.

If you need more results, use pagination with offsets.

Related Articles
Connecting to your Adobe Analytics account
Oncrawl bot: what URLs are known, crawled and analyzed?
How to use alerts for SEO monitoring in Oncrawl
Migrating from GA Universal Analytics to GA4 in Oncrawl
Introduction to the Oncrawl MCP Server
Did this answer your question?