Skip to main content
Glama
davidmosiah

Google Ads Intent MCP

by davidmosiah
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

⭐ If this agent-first tool helps your workflow, please star the repo. Stars make this tooling easier for other builders to discover and help Delx keep shipping open infrastructure. 🧱 Part of the Delx agent stack — 15 open-source MCP servers across body, reach and coordination.

Dry-run-first Google Ads search-term intent analyzer for agents. It helps Codex, Claude, Cursor, Hermes, OpenClaw and other MCP clients classify search terms, protect buyer intent and draft negative-keyword plans from CSV exports before any live account change.

Use it when an agent needs to reduce wasted spend without accidentally excluding buyer-intent queries.

HTTP (v2 stateless)

Not shipped. Optional FastMCP extra is stdio-only (same skip as delx-agent-utilities: no /health without extra Starlette routes). Dry-run analysis stays request-stateless over stdio.

Related MCP server: mcp-gads

Why It Exists

Google Ads cleanup is risky when agents act directly on accounts. This package makes the safe path the default:

  • analyze exported search-term CSVs locally

  • classify waste, buyer, research and competitor intent

  • draft negative-keyword plans without applying them

  • expose manifest, connection_status and privacy_audit before action tools

  • keep live mutation out of v0.1

Install

pipx install google-ads-intent-mcp

With MCP support:

pipx install "google-ads-intent-mcp[mcp]"

Published on PyPI: google-ads-intent-mcp. Release automation uses PyPI Trusted Publishing, so GitHub Actions can publish future versions without long-lived PyPI tokens. See docs/pypi-publishing.md.

CLI

google-ads-intent manifest --client codex
google-ads-intent doctor
google-ads-intent privacy-audit
google-ads-intent classify "free robux generator no verification"
google-ads-intent analyze-csv --csv examples/search_terms.csv
google-ads-intent plan-negatives --csv examples/search_terms.csv

Intent classification

The classifier is a deterministic, dependency-free heuristic with broad, cross-vertical signal coverage (ecommerce, B2B/SaaS, local services, health, finance, education and more) — not just gaming traffic. It sorts each search term into waste, buyer, research or competitor intent and protects converting queries from being flagged as negatives.

An optional LLM/embeddings-backed refinement path is available and is off by default. It requires no extra dependencies or API keys for normal use, and always falls back to the heuristic when no backend is configured:

# Opt in via flag (falls back to the heuristic if nothing is configured)
google-ads-intent --llm classify "crm software pricing"

# Or via environment variable
GOOGLE_ADS_INTENT_LLM=1 google-ads-intent analyze-csv --csv export.csv

To actually call a backend, set OPENAI_API_KEY (and optionally GOOGLE_ADS_INTENT_LLM_MODEL, default gpt-4o-mini) and install the openai package. Without those, --llm is a no-op that keeps the heuristic result. Each classification reports which path produced it via a source (heuristic or llm) field.

MCP

google-ads-intent-mcp

Hermes-style config:

mcp_servers:
  google_ads_intent:
    command: google-ads-intent-mcp
    args: []
    sampling:
      enabled: false

Recommended first calls:

  1. google_ads_connection_status

  2. google_ads_privacy_audit

  3. google_ads_analyze_search_terms

  4. google_ads_build_negative_plan

Agent Surfaces

Tool

Purpose

google_ads_agent_manifest

Install/runtime guidance for agent clients

google_ads_connection_status

CSV/API readiness without credentials

google_ads_privacy_audit

Dry-run, account and export boundaries

google_ads_classify_search_term

Single-query intent classification

google_ads_analyze_search_terms

Batch CSV-style analysis

google_ads_build_negative_plan

Dry-run negative keyword plan

Copy-Paste Agent Prompt

Use google-ads-intent-mcp. First call google_ads_connection_status and google_ads_privacy_audit.
Analyze the search terms, protect buyer/conversion queries, and return a dry-run negative plan only.

CSV Format

The parser accepts common exported columns such as:

  • search_term, Search term, Query

  • cost, Cost, cost_micros

  • clicks, Clicks

  • conversions, Conversions, Conv.

  • impressions, Impr., Impressions

Safety Model

  • CSV analysis is local.

  • Negative plans are dry-run only.

  • Buyer/conversion terms are protected from automatic exclusion.

  • OAuth tokens, developer tokens and account identifiers should stay in local environment/config files.

Development

python3 -m venv .venv
. .venv/bin/activate
pip install -e ".[dev]"
pytest
python -m compileall -q src

📧 Contact & Support

Available Tools

6 tools

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

  1. 6 tool updatesv0.1.1
    • First observedgoogle_ads_agent_manifest
    • First observedgoogle_ads_analyze_search_terms
    • First observedgoogle_ads_build_negative_plan
    • First observedgoogle_ads_classify_search_term
    • First observedgoogle_ads_connection_status
    • First observedgoogle_ads_privacy_audit
Deploy Server
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

TDQS

C2/5.0

Scored across 6 tools

Disambiguation4/5

Tool names suggest distinct purposes: analyzing search terms, classifying a single term, building negative plans, checking status, and auditing privacy. Potential confusion between 'analyze_search_terms' and 'classify_search_term' is mitigated by singular/plural distinction, but 'agent_manifest' is somewhat vague.

Naming Consistency4/5

All tools share the 'google_ads_' prefix and use snake_case. However, the pattern after the prefix varies: some use verb_noun (analyze_search_terms, classify_search_term, build_negative_plan) while others use noun (connection_status, privacy_audit) or noun_noun (agent_manifest).

Tool Count5/5

With 6 tools, the server is well-scoped for its apparent domain of search term analysis and negative keyword management. The count is neither too sparse nor excessive.

Completeness4/5

The tool set covers core operations for search term intent analysis and negative plan creation, plus connection and privacy checks. Missing are update/delete operations for negative plans or retrieval of existing plans, but the set is largely sufficient for its focused purpose.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • mcp-gads

    pijusz
    A
    license
    Not graded
    quality
    B
    maintenance
    Enables querying and managing Google Ads campaigns, keywords, assets, and more via natural language, with support for multiple MCP clients.
    88 npm
    1
    MIT

  • google-ads-mcp-plus

    monsieurgoodmood
    A
    license
    A
    quality
    B
    maintenance
    Extends the official Google Ads MCP with a safe write layer for creating paused-by-default Search campaigns and an account auditor, all running locally with no hosted dependencies.
    13
    1
    Apache 2.0

  • google-ads-write-mcp

    Luxand
    A
    license
    A
    quality
    A
    maintenance
    An MCP server that enables safe, audited mutation of Google Ads campaigns—creating ads, ad groups, keywords, and assets or adjusting budgets and statuses—with a dry-run default and an optional guarded remove operation, plus read-only Keyword Planner ideas.
    22
    1
    MIT