Artificial Intelligence1 min read

Llms.txt and MCP didn't move the needle

By · Published by Everything Blog

In short

The creator of Backpack for Laravel discovered that artificial intelligence agents rarely look for llms.txt files and struggle to formulate useful documentation queries via Model Context Protocol servers. AI development success instead relies on foundational choices, such as preserving a stable API across four years to ensure accurate public training data. Furthermore, having open-source repositories and installing readable, unobfuscated code directly into local project directories allows models to inspect real source files to solve developer problems.

Key points

  • Specialized artificial intelligence integrations like Model Context Protocol servers ofte…: Specialized artificial intelligence integrations like Model Context Protocol servers often fail because models struggle with search queries.
  • Providing open-source code and long-term public documentation creates superior training f…: Providing open-source code and long-term public documentation creates superior training foundations compared to custom AI context files.
  • Package creators should focus on delivering readable, unobfuscated local source code rath…: Package creators should focus on delivering readable, unobfuscated local source code rather than relying on external AI-fetching tools.

Sources checked by our analysis pipeline: 1

Two years ago, an admin panel for a simple website took me a day, maybe two - and I was proud of how fast I was. Now it takes me about...

Two years ago, an admin panel for a simple website took me a day, maybe two - and I was proud of how fast I was. Now it takes me about an hour or two.

The old day went like this. Install the package, generate the CRUDs, configure every field and column, wire up the relationships, click through everything to check it works. Thirty to sixty minutes per CRUD, easily. Now I describe what I want, the agent runs our generators, writes the configuration, writes the tests, and I review. A day of work fits in an hour. A week-long feature fits in a day.

I'm not the only one seeing this. In February 2026, a customer from New Zealand gave us a 10 out of 10 in our feedback form. The whole review was one line:

"Straightforward to use (particularly with Claude looking over my shoulder!)"

That's the…

Artificial Intelligence