Summary
Build a CLI entry point (google-ads-cli) alongside the existing MCP server for automated, scheduled ad management workflows — starting with daily search query review and negative keyword management.
Motivation
- MCP is great for interactive exploration but has token/context limits for large data
- Daily search query review is repetitive, rule-based, and schedulable — better suited to a CLI
- Both interfaces share the same client, config, and helpers
Architecture
Same repo, same package. Add cli.py entry point and workflows/ directory:
src/google_ads_mcp/
cli.py # CLI entry point (Typer/Click)
workflows/ # Multi-step automation logic
search_query_review.py
negative_keywords.py
[project.scripts]
google-ads-mcp = "google_ads_mcp.server:main"
google-ads-cli = "google_ads_mcp.cli:main"
First Workflow: Search Query Review
- Pull search terms report (configurable date range, cost threshold)
- Apply rules (known irrelevant patterns, zero-conversion filters)
- Optionally call Claude API for ambiguous term classification
- Export report (CSV)
- Auto-add obvious negatives, flag borderline ones for review
Example Usage
google-ads-cli search-query-review \
--account 6238255881 \
--days 7 \
--min-cost 5.00 \
--export report.csv \
--auto-negative
Labels
backlog, enhancement
Summary
Build a CLI entry point (
google-ads-cli) alongside the existing MCP server for automated, scheduled ad management workflows — starting with daily search query review and negative keyword management.Motivation
Architecture
Same repo, same package. Add
cli.pyentry point andworkflows/directory:First Workflow: Search Query Review
Example Usage
Labels
backlog, enhancement