Project status: stabilization, testing, and maintainer roadmap

[DeusData]

Hi everyone,

First of all: thank you for the incredible response to codebase-memory-mcp.

The project is still very young — around four months old — and over the past week it grew from roughly 3.5k stars to around 13k. That kind of growth is exciting, but it also means the project is now being tested against many more real-world codebases, platforms, repository sizes, and agent workflows than before.

Because of that, the next phase will intentionally focus on stabilization, stronger testing, and improving reliability before expanding the scope too aggressively.

The most valuable help right now would be:

• extending existing test cases
• adding regression tests for bugs that have already been found
• testing the tool against larger and more diverse repositories
• hunting for edge cases in parsing, indexing, graph queries, and memory usage
• improving platform-specific coverage, especially where behavior differs between Linux, macOS, and Windows
• contributing minimal reproductions that can be turned into automated tests

If you want to help with bug hunting, stabilization, or regression testing, please take a look at the existing umbrella tasks first. They are meant to collect related reports, known failure modes, linked PRs, reproduction cases, and open questions in one place.

The most relevant umbrella areas right now are:

• task: Stability — crashes, segfaults, and memory safety during indexing (14 issues) #390 — crash resistance and stability
• task: Silent index corruption / incomplete index reported as success (3 issues) #391 — silent corruption / incorrect analysis results
• task: Parsing & extraction quality — language/format coverage gaps #592 — parsing and extraction correctness
• task: Performance & memory on large repositories #593 — performance and memory behavior on large repositories
• task: Graph queries & trace_path resolution #594 — graph query and path resolution correctness
• 📌 Epics / Roadmap — umbrella task index #595 — overall umbrella index / project stabilization overview

Before starting larger work in one of these areas, please check the corresponding umbrella task and coordinate there when possible. This helps avoid duplicate work and makes it easier to turn bug reports into regression tests.

If you find a bug, the ideal contribution is not only a report, but also a minimal reproduction or a test case that captures the behavior. Extending existing tests — or proposing a dedicated new test class or test suite where it makes sense — is especially valuable right now.

I am also considering whether some areas would benefit from new dedicated test classes or more structured test suites. I still need to think through the best design for that, and I would like to discuss it with contributors before making larger changes to the test structure.

In parallel, I am working on a concept for bringing additional maintainers on board. The project has grown quickly, and I want to make sure that new maintainers can help in a way that protects quality, keeps the direction clear, and gives contributors a reliable review path. I still need to work out the details, and I am open to discussing what a good maintainer model should look like for this project.

Feature requests and new integrations are still welcome, but the immediate focus is on making the foundation more reliable: better tests, better reproductions, better coverage, and fewer surprises across real-world repositories.

Thank you again for the support, feedback, bug hunting, testing, and contributions. The goal now is to turn the recent momentum into a stronger and more reliable foundation for everyone using codebase-memory-mcp.

[ilyabrykau-orca]

@DeusData Just dropping by to say a massive thank you for codebase-memory-mcp, and huge congrats on crossing 13k stars!

I know maintaining a project that blows up this fast can be exhausting, so I really hope you're taking a moment to celebrate. As a thought: don't hesitate to lean heavily on your community right now. Empowering a few active users as core contributors to handle the day-to-day could be a total game-changer. You’d still keep the absolute final word on the strategy and vision, but with way less burnout.

If you let the community engine take on some of the load, this project is going to completely skyrocket and eclipse the alternatives (even Serena, especially if community-driven LSP editing gets added down the line).

Keep your head up, take care of yourself, and thank you again for building something so incredibly useful.
Best,

[DeusData]

Thank you, and yes, I am working on a concept to invite co-maintainers and give people more freedom. I don't want this project to be a solo show and this will also not be beneficial to it if I would go this route. Just needing to learn the basics to do it right as this is also obviously new to me (I guess OSS success is not something you can prepare for haha). I am very thankful for this experience and looking forward to how this evolves