I've recently been trying out and using the latest and greatest software development agents, including Windsurf and Cursor.

Before using those, I had been using aider for everyday practical SWE tasks, but aider is not very agentic --it's best at one-off, scoped programming tasks.

I did a quick test by putting aider in a ReAct loop using create_react_agent from langchain. To my surprise, it ended up working very well --it's able to solve programming tasks that Windsurf has failed at.

The agent has access to many tools such as ripgrep, fuzzy find, filesystem navigation, etc. One tool, in particular, is an expert tool where it can use a dedicated reasoning model, such as OpenAI's o1, to reason about logic, bugs, or complex planning tasks. It's been really interesting to see when and how it decides to use these tools.

The project does support plugging in any model you want to use, but so far I've found it to work best with the latest claude as the agent, and o1-preview (or o1 if you have access to it) for the expert reasoning model.

This is something I put together out of practical need, so if you have ideas, feedback, or especially PRs, those are welcome.

The project is licensed as Apache 2.0 (same as aider) and is intended to be a public good --our open tool so we aren't all dependent on proprietary tools and services to get our work done. The dream is that it will eventually work well on fully open/local models.

Cheers