ListFixer

ListFixer started as one AI-driven prospecting tool that you can use to find companies in your ICP.

You can for example upload a list of 10000 company websites, have AI analyse all of them and tell you which companies are B2B or B2C.

More similar tools were built, and the decision was made to turn it into a software platform that other people can pay to use.

Most of the tools are centred around bulk processing websites or lead data.

You upload data in the thousands.

Tell AI how you want to segment or split the data, and the job is processed.

It's an easy way to find the companies worth pitching your offer to, and leave out the rest.

Each tool is what we call an AI workflow, meaning there's one system prompt being reused over and over again with different context.

If you upload 1000 websites, each website is scraped, and the same AI model is given the same question about 1000 different websites.

Each workflow can have one or several steps.

If it has several steps, that means we make one call to an AI model, and then the output from that one is given to a second call to another AI model.

This is pretty easy to build.

The hard part is making a good system prompt that works and creating a good web scraper or research machine.

These kinds of repetitive automation workflows can usually be set up quickly.

In order to process thousands of websites with AI for free, we created an MCP connector.

ListFixer's tools normally use an API connection to Anthropic or OpenAI to use their AI models for the tools.

Processing jobs with the tools using the MCP, we can power the AI brain in the tools with our Claude Code subscription instead of by API connection.

This is roughly 20x cheaper.

Users can then choose if they want to run jobs using their own Claude Code subscription or run them through our API connection to Anthropic.

A nice trick to minimise costs of AI usage.