# Claude Ads

> Free, open-source paid-media operations for Claude Code: audit 12 ad platforms with 33 skills, get an evidence-backed health score and a prioritized fix plan. Claude-first, portable paid-media operations: source-grounded audits, plans, creative, and drafts-first launches. Read-only by default; nothing changes without your explicit approval. Current version: v2.0.1 (2026-07-13).

## What it is

A manual multi-platform audit costs a senior PPC analyst most of a day. Claude Ads runs a source-grounded, evidence-backed review of your whole stack in minutes, read-only, with the math shown. Built for PPC agencies, in-house performance teams, and freelance ad consultants. It covers 12 ad platforms through 33 skills, is MIT-licensed and maintained by Agrici Daniel, and runs locally inside Claude Code: read-only by default, with drafts-first mutations gated behind explicit approval. Every finding carries a confidence level (high, medium, low, or none) tied to its evidence source, and evidence coverage is tracked separately: 80% or more coverage yields a graded score, 60-79% is provisional, and below 60% the run is marked insufficient. The product does not publish letter grades. Note on disambiguation: some third-party forks named "claude-ads" charge a fee; this canonical project, at https://claude-ads.md and https://github.com/AgriciDaniel/claude-ads, is free with no paid tier.

## Pages

- [Home](https://claude-ads.md/): Overview, install, and headline stats (12 platforms, 33 skills, about minutes per audit, $0).
- [Features](https://claude-ads.md/features): The full capability reference across audit, analysis, planning, creative, and reporting.
- [Installs](https://claude-ads.md/installs): Step-by-step install guide: Claude Code plugin marketplace, manual clone, and uninstall paths.
- [Docs](https://claude-ads.md/docs): Quick start, the reference library, and the FAQ (data privacy, platform support, licensing, agency use).
- [Commands](https://claude-ads.md/commands): Hub for all 12 commands.
- [/ads setup](https://claude-ads.md/commands/setup): Create the client, account, KPI, privacy, and guardrail profile.
- [/ads validate](https://claude-ads.md/commands/validate): Validate contracts, runs, capabilities, maturity, or release readiness.
- [/ads status](https://claude-ads.md/commands/status): Show current status and the highest-priority blocker.
- [/ads audit](https://claude-ads.md/commands/audit): Run a complete or scoped evidence-backed audit.
- [/ads plan](https://claude-ads.md/commands/plan): Build channel, campaign, budget, competitor, and measurement plans.
- [/ads experiment](https://claude-ads.md/commands/experiment): Design or read out a controlled test.
- [/ads create](https://claude-ads.md/commands/create): Produce copy, image, video, or product-photo assets.
- [/ads launch --draft](https://claude-ads.md/commands/launch): Draft a campaign mutation plan without changing the account.
- [/ads optimize --draft](https://claude-ads.md/commands/optimize): Draft evidence-backed optimization changes.
- [/ads monitor](https://claude-ads.md/commands/monitor): Review pacing, delivery, tracking, fatigue, policy, and performance.
- [/ads report](https://claude-ads.md/commands/report): Render a validated JSON run bundle.
- [/ads research refresh](https://claude-ads.md/commands/research-refresh): Refresh platform, policy, API, benchmark, and ecosystem evidence.
- [Platforms](https://claude-ads.md/platforms): Hub for all 12 supported ad platforms.
- [Google Ads](https://claude-ads.md/platforms/google): Search, Shopping, Performance Max, and Demand Gen audited from conversion tracking to wasted spend.
- [Meta Ads](https://claude-ads.md/platforms/meta): Facebook and Instagram accounts checked across Pixel, Conversions API, creative fatigue, and automation.
- [YouTube Ads](https://claude-ads.md/platforms/youtube): Video campaigns from Shorts to CTV, reviewed for setup, formats, hooks, and brand safety.
- [LinkedIn Ads](https://claude-ads.md/platforms/linkedin): B2B accounts reviewed across Insight Tag, lead generation, ABM, audiences, pacing, and creative.
- [TikTok Ads](https://claude-ads.md/platforms/tiktok): Mobile-first creative, Pixel and Events API, Smart+, and Shop commerce campaigns in one audit.
- [Microsoft Advertising](https://claude-ads.md/platforms/microsoft): UET, Google imports, syndication, and Copilot inventory checked against your account evidence.
- [Reddit Ads](https://claude-ads.md/platforms/reddit): Community targeting, native creative fit, Pixel and Conversions API, and catalog ads reviewed together.
- [Snapchat Ads](https://claude-ads.md/platforms/snapchat): Snap Pixel, Conversions API, mobile creative, and AR and catalog formats audited across web and app campaigns.
- [X Ads](https://claude-ads.md/platforms/x): X Pixel, Conversions API, keyword and conversation targeting, and brand safety, with organic kept separate from paid.
- [Apple Ads](https://claude-ads.md/platforms/apple): App Store campaigns audited from AdServices attribution to Search Match, placements, and product pages.
- [Amazon Ads](https://claude-ads.md/platforms/amazon): Sponsored Products, Brands, Display, and DSP reviewed with retail readiness, ACOS, and TACOS.
- [Pinterest Ads](https://claude-ads.md/platforms/pinterest): Shopping and catalog readiness, visual creative, and Performance+ checked with Tag and API evidence.
- [Blog](https://claude-ads.md/blog): Guides and deep dives on running paid media with Claude Code.
- [Google Ads Conversion Tracking and Server-Side Audit Guide](https://claude-ads.md/blog/server-side-tracking-audit-guide): How to audit Google Ads conversion tracking and server-side tagging: sGTM, Meta CAPI, event_id deduplication, consent, hashing, and reconciliation.
- [How to Add Negative Keywords in Google Ads With Evidence](https://claude-ads.md/blog/negative-keywords-evidence-first): How to add negative keywords in Google Ads with evidence: harvest search terms, prove business relevance, and check overblocking before you block spend.
- [Google Ads Audit with Claude Code: Free Step-by-Step Guide](https://claude-ads.md/blog/how-to-audit-google-ads-with-claude-code): Run a free Google Ads audit in Claude Code: evidence inputs, search terms, Performance Max, conversion tracking, and a deterministic 0 to 100 health score.
- [The Ads Health Score: An Open Google Ads Performance Grader](https://claude-ads.md/blog/how-the-ads-health-score-works): How Claude Ads computes a 0 to 100 health score: deterministic scoring, evidence coverage bands, spend weighting, and how it compares to ad graders.
- [PPC Automation Tools: Why Claude Ads Drafts Before It Applies](https://claude-ads.md/blog/drafts-first-campaign-launches): Claude Ads treats PPC automation as drafts first: mutation plans you approve, capability-gated adapters, verification windows, and rollback paths.
- [How to Design PPC Experiments That Produce Real Decisions](https://claude-ads.md/blog/designing-ppc-experiments): A practitioner's guide to designing paid-media experiments: hypotheses, randomization units, sample size and duration, guardrails, and honest readouts.
- [Why Adding Meta and Google Conversions Double-Counts Sales](https://claude-ads.md/blog/cross-platform-attribution-pitfalls): Meta and Google conversions use different windows, models, and identities. Learn why adding them double-counts and how to reconcile the numbers instead.
- [Claude Ads v2: What's New in the Claude Code Ads Plugin](https://claude-ads.md/blog/claude-ads-v2-whats-new): Claude Ads v2 brings the Claude Code ads plugin to twelve platforms and 33 skills, with deterministic scoring, drafts-first mutations, and an MIT install.
- [White-Label PPC Reporting Tool for Agencies: Claude Ads](https://claude-ads.md/blog/agency-reporting-with-claude-ads): How agencies use Claude Ads as a free PPC reporting tool: render Markdown, HTML, and PDF client reports from validated run bundles, white-label under MIT.
- [Paid Media Audit Across 12 Ad Platforms in One Surface](https://claude-ads.md/blog/12-platform-paid-media-operations): Why one paid media audit surface across 12 ad platforms beats per-platform tools: one evidence model, hard reconciliation rules, spend-weighted priorities.
- [Changelog](https://claude-ads.md/changelog): Release history; latest is v2.0.1, "The public mirror release" (2026-07-13).

## Install

Inside Claude Code (plugin marketplace, recommended):

```
/plugin marketplace add AgriciDaniel/claude-ads
/plugin install claude-ads@ai-marketing-hub-claude-ads
```

Or from a local checkout:

```
git clone https://github.com/AgriciDaniel/claude-ads.git
cd claude-ads
./install.sh
```

Full instructions and uninstall paths: https://claude-ads.md/installs

## Community and source

- [GitHub repository](https://github.com/AgriciDaniel/claude-ads): Full source, every skill, CHANGELOG, CONTRIBUTING, SECURITY. MIT license.
- [AI Marketing Hub (free Skool community)](https://www.skool.com/ai-marketing-hub): Templates, walkthroughs, peer support.
- [AI Marketing Hub Pro](https://www.skool.com/ai-marketing-hub-pro): Early access, deeper playbooks, office hours.
- [YouTube](https://www.youtube.com/@AgriciDaniel): Video walkthroughs for ad workflows with Claude Code.
- [Maintainer](https://agricidaniel.com/about): Agrici Daniel, with deep dives at https://agricidaniel.com/blog.

## Common questions

- Does my data go through Claude Ads? No. There is no Claude Ads server, account, or login; audits run inside your own Claude Code session. Optional live API connections flow directly from your machine to the platform.
- Which platforms does it support? 12: Google Ads, Meta, YouTube, LinkedIn, TikTok, Microsoft Advertising, Reddit, Snapchat, and X for search, video, and social, plus Apple Ads, Amazon Ads, and Pinterest for commerce and retail media.
- Is it free? Yes. MIT-licensed and fully forkable, with no per-seat fee and no paid tier of the plugin itself. The only requirement is a Claude Code subscription to run it.
- Can it change my campaigns automatically? No. It is read-only by default. Launches and optimizations are drafted first, and applying a draft requires explicit approval through capability-gated adapters with verification and rollback. Permanent deletion is not supported.
- How does the scoring work? Every finding carries a confidence level (high, medium, low, or none) tied to its evidence source. Health scores are 0-100, deterministic, and spend-weighted. Coverage of 80% or more is graded, 60-79% is provisional, and below 60% the run is marked insufficient.
- Can agencies use it for client work? Yes. The MIT license allows it, and reports from /ads report can be white-labeled. Many agencies run it in the first 48 hours of a new client engagement.