AdCopilot

Google Ads: find keyword ideas

get_keyword_ideas

Read-onlyIdempotent

Finds keyword ideas with Google's search volume, competition and top-of-page bid range.

Ideas grow from keywords, a landing page or a whole site: pass keywords
(1 to 20), page_url, keywords with page_url, or site alone. Figures are
Google's estimates from the last 12 months; avg_monthly_searches is null
where Google reports too few searches to show. A new account, or one
that has never spent, can get ranges or no ideas. It reads only and
saves nothing.

Each call is one request to Google and returns one page of at most
`limit` ideas; next_page_token asks for the next page. Google accepts
about one keyword planning request per second per account, so a request
within a second of the last one for the same account waits for the rest
of that second.

Args:
    customer_id: The Google Ads customer ID (numbers only, no hyphens).
    keywords: Seed keywords or phrases, 1 to 20.
    page_url: A page whose content seeds the ideas (http:// or https://).
    site: A domain whose pages seed the ideas, e.g. "example.com". Not
        combined with keywords or page_url.
    geo_target_ids: Location ids (geo target constants, e.g. 2356 for
        India; search geo_target_constant to find one), at most 10.
        Default: every location.
    language_id: Language id (language constant). Default 1000, English.
    network: GOOGLE_SEARCH (default) or GOOGLE_SEARCH_AND_PARTNERS.
    include_adult_keywords: Include adult ideas. Default false.
    include_monthly_trend: Add each idea's searches for the last 12
        months.
    include_average_cpc: Add the average cost per click. Default true.
    limit: Ideas on this page, 1 to 200. Default 50.
    page_token: next_page_token from the call before, for the next page
        of the same request.

Returns:
    ideas (text, avg_monthly_searches, competition LOW/MEDIUM/HIGH,
    competition_index 0-100, low_top_of_page_bid and high_top_of_page_bid
    in the account's currency, average_cpc, close_variants), returned,
    total_size, truncated, next_page_token, query and notes.

Tools revision: 5c9bea8f.

Input Schema

TableJSON Schema

Name	Required	Default
`site`	No
`limit`	No
`network`	No	GOOGLE_SEARCH
`keywords`	No
`page_url`	No
`page_token`	No
`customer_id`	Yes
`language_id`	No
`geo_target_ids`	No
`include_average_cpc`	No
`include_monthly_trend`	No
`include_adult_keywords`	No

Output Schema

TableJSON Schema

Name	Required	Description	Default
`result`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

First observedOct 23, 2026

TDQS

A4.6/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Goes well past the readOnlyHint/idempotentHint annotations with substantive behavior: figures are Google's 12-month estimates, avg_monthly_searches is null when volume is too low, new accounts can return ranges or nothing, and each call is a single upstream request returning one page. The rate-limit paragraph (about one keyword planning request per second per account, with requests inside that second waiting) is exactly the kind of latency behavior an agent needs and cannot get from annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded: the first line states purpose and outputs, then seeding modes, then data caveats, then pagination/rate limits, with a clean Args/Returns block. Every sentence supplies information an agent would otherwise have to guess at; there is no filler or restatement of the tool name.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 12-parameter, open-world, rate-limited tool this is complete: it covers seeding inputs, defaults, pagination, quota pacing, data-quality caveats, and account-state edge cases, and even names the return fields despite an output schema already existing. Nothing an agent needs to construct a correct call or interpret a partial result is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so the description carries the full burden — and it does: every one of the 12 parameters is documented with format, bounds, and defaults (customer_id numbers-only, keywords 1-20, at most 10 geo ids with a discovery pointer, language default 1000, limit 1-200 default 50, page_token sourced from next_page_token). This compensates completely for the empty schema descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a specific verb and resource ('Finds keyword ideas') plus the payload it delivers (search volume, competition, top-of-page bid range), and the seeding-modes sentence makes clear this is a discovery/generation tool. It does not explicitly contrast itself with close siblings such as get_keyword_metrics, get_keyword_forecast, or keyword_opportunities, so the differentiation is inferred rather than stated.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives concrete selection rules among its own modes (keywords 1-20, page_url, keywords+page_url, or site alone) and a hard exclusion ('site ... Not combined with keywords or page_url'). It also warns when results may be empty (new or never-spent accounts). What it lacks is guidance on when to reach for this versus sibling keyword tools, which is exactly the ambiguity an agent faces at selection time.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All AdCopilot tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?