RepoLens - Turn repositories into AI-ready context packs.
by•
RepoLens is a CLI-based tool that turns your codebase into a single AI-ready package.
It helps developers:
* Analyze large codebases (currently supporting 9 programming languages)
* Generate structured AI-ready context packs
* Optimize token usage for LLMs
* Preserve repository structure and relationships
* Redact secrets (API keys, passwords, and other sensitive data)
* Integrate seamlessly into developer workflows
Replies
The secret redaction feature caught me off guard in a good way, actually caught an old AWS key I forgot was in a config file. Token optimization makes feeding repos to Claude way less painful too.
@alaykuw Thank you! That is exactly one of the problems I wanted RepoLens to help with. It's surprisingly easy for old credential to linger in reporsitories overtime, especially in config files.
I am also glad the token optimization helped in making it easier to work with Claude. If you have any ideas for improving the workflow or additional features you would like to see in RepoLens, I would love to hear them.
Love the secret redaction feature, that alone makes it worth checking out. One thing I'd want to see is configurable output formats beyond the default context pack, like a markdown summary or JSON schema, so I can pipe results straight into different LLMs or docs without extra conversion steps.
@erdoanv91s Thank you for the sugetion Erdogan! RepoLens currently supports both the .pack and JSON formats. But I completely agree that additional formats like Markdown summaries would make it easier to integrate with different LLM and document workflows.
You can already generate JSON output with:
Markdown export is a great idea and is something that I will consider in a future release. Thank you for the feedback and for checking out RepoLens.
love the secret redaction angle, that's such a real pain point when feeding code to llms. one thing that would make this a no-brainer for me is a `--diff` flag that only packages changed files since the last commit, would save a ton of tokens during iterative debugging sessions.
@kemaltzly Thank you Kemal! I am glad the secret redaction feature resonated with you - that was one of the first problems I wanted RepoLens to solve.
Packing only the files changed since the last commit with a --diff flag is a fantastic idea. It makes iterative debugging and code reviews much more token-efficient especially while working with large repositories.
Something like Repolens.jar build --diff would fit nicely with the existing CLI. I'll add it to the roadmap. Thanks for the suggestion!
The secret redaction actually caught an API key in a test repo I forgot was there. Token optimization is a nice touch for fitting bigger codebases into context windows.
@aysunakkulkzhi Tha is awesome to hear Aysun! Catching forgotten secrets befor getting sent to an LLM is exactly why I built the secret redaction feature.
And I am glad the token optimization is proving to be useful too - large repositories can hit context limits quickly. Thank you for checking RepoLens on a real project and sharing your experience.
Tried it on a messy Node project and the secret redaction actually caught a few API keys I'd forgotten were floating around in old config files. Token optimization is a nice touch too.
@savadgcs Thanks for giving RepoLens a try, Savas! That is exactly the kind of issue I built RepoLens to help solve - catching forgotten API keys and secrets lingering in repositories unnoticed.
I'm also glad you got the token optimization feature helpful. The goal is to make working with large codebases practical without wasting valuable context window space.
I really appreciate you sharing your experience!
The secret redaction step is a really thoughtful touch, especially for anyone planning to actually feed this into an LLM without leaking credentials. Solid execution.
@sleymanenseiwb Thank you for checking out RepoLens, Suleyman. That is exactly the concern I wanted to address - making it safer for developers to use LLMs without accidentally exposing credentials or sensitive information.
Secret redaction is a small step, but an important one when AI assisted developement becomes part of the workflow. Really appreciate your feedback!
Tried it on a messy monorepo and the secret redaction alone saved me from a dumb mistake. Token packing feels tighter than the other tools I've poked at.
@azadxpsa Thanks for trying RepoLens, Azad. I'm glad the secret redaction helped you avoid that mistake - that is exactly the kind of problem I wanted to prevent.
And I appreciate the feedback on the token packing. Making the generated context as focused and efficient as possible is a big part of the goal, especially for large codebases and monrepos.
Thanks for sharing your experience!
The secret redaction alone is worth the install, especially how it seems to preserve the file relationships while scanning. Nice to see a CLI tool that actually respects developer workflow instead of trying to bolt on another UI.
@nurtenwmvt Thanks for the kind words, Nurten! Preserving the repository structure and developer workflow was a big design goal from the beginning. I'm glad the secret redaction feature has been useful.
Really appreciate you taking the time to share your feedback!
Love the idea of packaging repos for LLM context. One thing that would make this a daily driver for me is adding incremental updates, so it only re-processes changed files instead of rebuilding the whole context pack every time. Would save a lot of time on bigger repos.
@cengizgalerisi Thanks Cengiz! That's a great suggestion. Incremental updates are definitely something I've been thinking about.
A --diff or incremental mode that only processes changed files since the last run would nicely fit with the existing workflow. I'll add it to the roadmap - really appreciate the feedback!