Hello, world
We are starting a blog. Here we will share how we design a13n and what we think about agents and building software.
This is the first post on the a13n blog.
a13n is the open-source foundation for AI agents. Harness is the foundation that runs an agent, Harness UI is a playground for trying agents, and Service operates managed agents on cloud. Most of what we decide while building them is recorded in issues, pull requests, and specifications. Those records keep what we decided. They rarely keep why one design won over another, or what we learned on the way.
This blog is where we write that down:
- Designs. A problem we had to solve, the options we weighed, and why we chose the one we shipped.
- Engineering. The mechanisms inside Harness and Service, with the numbers that shaped them.
- Thinking. What we believe about agents, about models that keep getting better, and about building software now that agents write much of it.
Agents and engineering will come up most often, but they are not the only subjects.
The code is on GitHub. If a post raises a question, open an issue there.