claude-ads executes source-grounded paid-media operations and audits across multiple ad platforms. It turns authorized exports or account reads into versioned JSON reports covering audits, campaign plans, creative workflows, monitoring, and experiments.
Project overview
claude-ads enforces read-only adapters by default, requires explicit approval gates before any write operations, and produces source-grounded audits with deterministic scoring across 12 paid-media platforms.
Project type
AI Coding
Deployment
Refer to project documentation
License
MIT
Best for
Enterprise teams that need source-grounded paid-media audits, campaign planning, creative generation, and monitoring with deterministic scoring and read-only safety gates.
Key capabilities
Audits paid-media accounts with dated evidence and explicit confidence across 12 platforms.
Plans campaigns, channels, budgets, measurement, and experiments.
Creates copy, image, video, and product-photo briefs and assets.
Monitors pacing, delivery, tracking, fatigue, policy, and performance.
Produces versioned JSON, then renders Markdown, HTML, and optional PDF.
Drafts safe account changes without applying them by default.
Separates health, evidence coverage, and regulatory exposure with a deterministic scoring system.
Limitations and risks
Permanent deletion is not supported in v2.
When account-defined ceilings are missing, no write operation is performed.
Getting started
Clone the repository and run the install script without coding: git clone https://github.com/AgriciDaniel/claude-ads.git, cd claude-ads, then bash install.sh --source=local. Alternatively, it can be installed directly via the Claude Code plugin flow.