Metricanic tracks clicks, conversions, events and costs from all your ad sources (Google Ads, Meta, TikTok, ad networks) in one account. Ask AI which ads lose money, and let it act on the answer: the built-in MCP server lets AI read your reports, re-route visitors by country, device or OS, and change campaign settings for you.
What became possible in your product with Astra that was not practical before?
Maker
Astra changed what I can realistically build into Metricanic.
The biggest example is MCP. With Astra, I finally made Metricanic’s MCP server write-capable, so AI can now actually change campaigns instead of only reading data.
Now I can take on much bigger changes, move faster, and catch more problems before they make it into production.
Report
Maker
📌
Hey Product Hunt 👋
I'm Viacheslav, the maker of Metricanic. I've been running paid traffic for my own projects since 2018, and for years my routine looked the same: launch a campaign, set up A/B tests, pull reports, pick a winner (sometimes too early), optimize, repeat.
When AI models got really good at reasoning, I wondered what would happen if one of them got direct access to all my campaign data in one place. So I built a tracker with a built-in MCP server and connected Claude and ChatGPT to it.
It worked better than I expected. In one campaign, the AI noticed that iOS users on Chrome from the UK converted 3x better than Android users from the UK on the same landing page. I'd never have spotted that three-way split on my own. I shifted bids toward that segment and scaled the campaign 5x at a lower cost per conversion. It also suggested which A/B tests to run and flagged campaigns that were simply burning money.
What Metricanic does:
• Tracks clicks, conversions, events and costs from all your ad sources in one place
• Routes visitors by country, device, OS and 100+ more dimensions
• Lets Claude, ChatGPT or any MCP-compatible AI read your reports and change routing and campaign settings from a chat
I'd love your feedback!
Astra changed what I can realistically build into Metricanic.
The biggest example is MCP. With Astra, I finally made Metricanic’s MCP server write-capable, so AI can now actually change campaigns instead of only reading data.
Now I can take on much bigger changes, move faster, and catch more problems before they make it into production.