Hosted Google Ads MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Scored across 4 tools
execute_action and query_data have near-identical descriptions (both are 'Step 3 (FINAL)' of a workflow requiring query_contexts first), differing only by 'action' vs 'data', so an agent cannot confidently tell read from write or mutation from report. query_contexts adds further overlap since it gates both of the others. Only prompts_menu has a clearly distinct purpose.
Two tools use verb_noun (execute_action, query_data), one uses verb_noun with a plural object (query_contexts), and prompts_menu is a noun_noun outlier. The pattern is readable but not uniform, and query_data/query_contexts are lexically close enough to be confused at the name level.
Four tools is thin, and two of them (execute_action, query_data) are effectively generic dispatchers that hide the real operation surface behind a 'request' parameter. The required two-step query_contexts preamble means much of the tool budget is spent on protocol plumbing rather than capability.
Read and write (execute_action) paths plus prompt discovery appear covered, but the actual operations are opaque—everything funnels through generic dispatchers, so an agent must make two dependent calls before knowing what is even possible. There is no discoverable surface for account/brand management beyond read-only context lookups.
Execute Google Ads action (Step 3 (FINAL) of workflow).
REQUIRED: Complete these steps first using query_contexts:
request='accounts' - Get workspace_id, brand_id, account_id AND available_requests
request='actions_details' with actions=['xxxxxx'] - Get detailed body structure for specific action(s) (use actions from available_requests)
IMPORTANT: Always call step 2 to see the exact body structure before executing.
Rely on info from steps 1-2 to construct the action body. Follow the schema exactly.
Inputs: body (required). Returns: object
Helper for planning InsightfulPipe ads data queries.
Step 1: request='accounts' – Retrieve workspace_id, brand_id, ad_account_id, other source metadata, AND available_requests (all tables you can query).
Step 2: request='actions_details' with actions=['table_a', 'table_b'] – REQUIRED before querying. Returns the exact request body, required identifiers, fields, time ranges, and notes for each action.
Optional: request='available_actions' – Standalone call to list all available tables/actions (already included in accounts response).
Optional: request='brands' with brand_id + workspace_id to pull brand guidelines.
Inputs: request (required), actions, brand_id, platform, +1 more. Returns: object
Query Google Ads data (Step 3 (FINAL) of workflow).
REQUIRED: Complete these steps first using query_contexts:
request='accounts' - Get workspace_id, brand_id, account_id AND available_requests
request='actions_details' with actions=['xxxxxx'] - Get detailed body structure for specific action(s) (use actions from available_requests)
IMPORTANT: Always call step 2 to see the exact body structure before querying.
Rely on info from steps 1-2 to construct the query body. Follow the schema exactly.
Inputs: body (required). Returns: object
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Hosted Meta Ads MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Hosted OpenAI Ads MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Hosted TikTok Ads MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Hosted DataForSEO MCP server for Claude, ChatGPT and Cursor, with each action switched per account.
Add one secure layer between your agents and this server.
