what is adspilot fr?

AdsPilot is a tool for automating and optimizing Google Ads campaigns, built around the idea that an AI, not a web dashboard, is the interface. Instead of logging into a hosted platform, users install AdsPilot locally, connect their own Google Ads account through a browser based authorization step, and then talk to an AI in plain language to select offers and keywords, build and adjust ad campaigns, filter unwanted traffic, and manage multiple accounts. The README explains that traditional advertising software needs login systems, subscription billing, and multi tenant permissions because the platform is doing the work on your behalf. AdsPilot flips this: the AI does the work, credentials and OAuth tokens stay on the user's own machine, and nothing is sent to a remote server, so those hosted-platform pieces are simply not needed. Four core capabilities are described. First, a pluggable affiliate provider interface lets the AI choose offers and write ad copy from structured affiliate data, with new affiliate networks added without touching the core system. Second, a keyword provider interface abstracts keyword data sources, starting with the Google Ads API and falling back to CSV import if API access is not yet approved. Third, real traffic is filtered through four layers, ranging from millisecond level fingerprint and rate checks to slower behavioral signals and offline conversion feedback, explicitly described as filtering out fake traffic rather than generating any. Fourth, multi account management uses Google Ads MCC alongside a fingerprint browser provider to keep account sessions isolated. The system is built as a set of independent Go service modules connected through a Go workspace, covering areas like the Google Ads API layer, an AI abstraction layer, site and keyword ranking, recommendations, and a console backend, with several modules such as affiliate integration and anti fraud filtering still planned. The project states firm rules against generating fake traffic or cloaking content differently for search engine crawlers versus real users, and that credentials are never stored in code, only in environment variables. The project is released under the MIT license and is still under active development.