What is Claude Ads?
Claude Ads is a free, open-source skill for agencies, consultants, and in-house marketing teams working with paid advertising. It accepts CSV or JSON exports, screenshots, and account summaries, then reports findings and missing evidence across Google, Meta, YouTube, LinkedIn, TikTok, Microsoft, Apple, Amazon, Reddit, Pinterest, Snapchat, and X. Its workflows cover audits, budget and measurement planning, creative production, monitoring, optimization drafts, and experiment design.
Claude Code is the verified runtime. Installation uses its native plugin flow or reviewed local skill files with a Python core. The canonical output is versioned JSON, with Markdown, HTML, and conditional PDF renderings. An optional local companion, Claude Ads Cockpit, exposes audit evidence, missing inputs, comparisons, and report receipts without being required by the main skill.
The project is a development candidate, not release-ready. All account-health scoring profiles are disabled, so audits provide findings, unknowns, and evidence coverage rather than health scores. Account adapters are read-only by default; changes remain drafts unless operation-specific capabilities, owner approval, limits, rollback, and verification requirements are satisfied.
Key capabilities
- Evidence-backed audits Classifies checks as pass, fail, unknown, or not applicable and associates findings with supporting evidence.
- Campaign planning Produces channel, campaign, budget, competitor, and measurement plans from the account context.
- Creative workflows Produces advertising copy, image, video, and product-photo briefs and assets.
- Monitoring and optimization Reviews pacing, delivery, tracking, fatigue, policy, and performance, with improvements drafted for review.
- Experiment support Designs controlled tests or prepares readouts of completed experiments.
- Structured reporting Saves validated JSON run bundles and renders Markdown, HTML, or PDF reports when dependencies are available.
Where Claude Ads fits
Claude Ads fits into account reviews and campaign planning where a marketing team can supply dated account evidence and review the resulting recommendations. Agencies and consultants can use saved runs to inspect findings and missing inputs. Cockpit can compare compatible runs, but those comparisons do not establish causality or advertising uplift.
Before deployment, operators should verify the enabled capabilities for each intended platform operation, input completeness, privacy rules, and account approval limits. Claude Code is the only verified host; other declared hosts remain experimental. Browser capture requires a separately installed Playwright browser payload, while PDF rendering requires WeasyPrint and native libraries. Cockpit requires explicit trusted core and Python paths, and its native host acceptance remains pending.
Install
/plugin marketplace add agricidaniel/claude-ads
/plugin install claude-ads@ai-marketing-hub-claude-adsReviews of Claude Ads
Be the first to share your experience with Claude Ads.
Write a review
Loading review form…