Critter Stack Roadmap for March 2026

It’s only a month since I’ve written an update on the Critter Stack roadmap, but it’s maybe worth some time on my part to update what I think the roadmap is now. The biggest change is the utter dominance of AI in the software development discourse and the fact that Claude usage has allowed us to chew through a shocking amount of backlog in the past 6 weeks. That’s probably also changed my own thinking about what should be next throughout this year.

First, some updates on what’s been added to the Critter Stack in just the last month:

By the time you read this, we may very well have Polecat 1.0 out as well.

Short Term

The short term priority for myself and JasperFx Software is to deliver the CritterWatch MVP in a usable form by the end of March.

Marten, Wolverine, and even Polecat have no major new features planned for the short term and I think they will only get tactical releases for bug fixes and JasperFx client requests for a little while. And let me tell you, it feels *weird* to say that, but we’ve blown through a tremendous amount of the backlog so far in 2026.

Medium Term

  • Enhance CritterWatch until it’s the best in class monitoring tool for asynchronous messaging and event sourcing. Part of that will probably be adding quite a bit more functionality for development time as well.
  • For a JasperFx Software client, we’re doing PoC work on scaling Marten to be able to handle having several hundred billion events in a single system. I’m going to assume that this PoC will probably lead to enhancements in both Marten and Wolverine!
  • We’ll finally add some direct support to Marten for the PostGIS PostgreSQL extension
  • I’m a little curious to try to use the hstore extension with Marten as a possible way to optimize our new DCB support
  • Play with Pgvector and TimescaleDb in combination with Marten as some kind of vague “how can we say that Marten is even more awesome for AI?”
  • There’s going to be a new wave of releases later this year for Marten 9.0, Wolverine 6.0, and Polecat 2.0 that will mostly about performance optimizations and especially finding ways to optimize the cold start time of applications using these tools.
  • Babu and I (really all Babu so far) are going to be building a set of AI skills for using the Critter Stack tools that will be curated in a GitHub repository and available to JasperFx Software clients. I do not know what the full impact of AI tools are really going to be on software development, but I personally want to plan for the worst case that AI tools plus LLM-friendly documentation drastically reduces the demand for consulting and try to belatedly pivot JasperFx Software to being at least partially a product company.
  • Build tooling for spec driven development using the Critter Stack. I don’t have any details beyond “hey, wouldn’t that be cool?”. My initial thought is to play with Gherkin specifications that generates “best practices” Critter Stack code with the accompanying automated tests to boot.
  • One way or another, we’ll be building MCP support into the Critter Stack, but again, I don’t know anything more than “hey, wouldn’t that be cool?”

Long Term

Profit?

I’m playing with the idea of completely rebooting Storyteller as a new spec driven development tool. I have the Nuget rights to the “Storyteller” name and graphics from Khalid (a necessary requirement for any successful effort on my part), and I’ve always wanted to go back to it some day.

Leave a comment