📣 Calling All Fork Contributors - We Want Your Feedback!

[mikejmorgan-ai]

Hey Cortex Community! 👋

We noticed many of you have forked and are actively working with Cortex Linux. We would love to hear from you!

Your feedback is incredibly valuable as we shape the future of this project.

---

Questions We'd Love Answered:

1. Did the MVP work for you?

• Were you able to get Cortex running?
• Any installation or setup issues?
• What environment are you running it on?

2. What would you improve?

• Missing features you need?
• Documentation gaps?
• Code quality concerns?
• Performance issues?

3. What are you building?

• We're curious what use cases you're exploring
• Any interesting integrations?

4. Would you contribute upstream?

• Have changes you'd like to submit as PRs?
• Found bugs you've fixed?
• Written tests or docs?

---

How to Respond:

• Reply directly to this discussion (preferred)
• Open an issue with the feedback label
• Submit a PR with improvements

---

Thanks for being part of the Cortex community! Every piece of feedback helps us build something better. 🚀

— The Cortex Team

[mikejmorgan-ai]

📢 Tagging Active Fork Contributors

We'd especially love to hear from those of you who have been actively working with Cortex:

@MChandrahas @gyanu2507 @kr16h @yaya1738 @hyaku0121 @JohnQMorgan @kishu279 @pavanimanchala53 @danishirfan21 @AlexanderLuzDH @shalinibhavi525-sudo @aliraza556 @chandrapratamar @Sahilbhatane @dhvll

Your feedback would be incredibly valuable! 🙏

[lu11y0]

Hi Mike - thanks for asking! Here’s my feedback (keeping it simple and clear).

---

1) Did the MVP work for you?

• Short: Yes, but I faced some issues.
• Was I able to run it? Yes, on Ubuntu (latest).
• Main issue: API key setup.
  • I tried setting API keys using environment variables and also entered them again in the wizard, but it still failed.
  • In the wizard, I was asked to choose an API provider. I selected OpenAI, set the key, but it still failed and showed an error like invalid Anthropic key or something similar. This was confusing.
  • I’m not sure if this was a mistake from my side or an app-side issue, but clearer error messages and validation would really help.
• Other issues:
  • pyyaml seemed missing as a dependency - I got an error on the first run because of that.

---

2) What would you improve?

• Installer / wizard
  • Make API key errors clearer (which key failed and why).
  • Show exact next steps (how to set env vars, where the config file is located).
  • Option to reset or reconfigure API keys easily.
• Terminal output
  • Improve UI using rich (colors, boxes, spinners). I’m working on this in issue Output Polish: Rich Formatting with Colors, Boxes, Spinners #242.
  • Make sure output still works properly when piped.
• Code quality
  • Avoid putting large logic directly in cli.py.
  • Better to keep functions in separate files and call them from cli.py.
  • Add basic linting and more tests, especially around config and setup.
• Debugging
  • When API keys fail, show helpful (non-sensitive) hints.

---

3) What am I building / using Cortex for?

• I implemented pre-built package stacks ([MVP] Pre-built package stacks (ML, Web, DevOps) #252).
• I’m currently working on cortex doctor (cortex doctor: System Health Check #245) and terminal output polish (Output Polish: Rich Formatting with Colors, Boxes, Spinners #242).
• Planning to work on:
  • stdin piping support (Stdin Piping Support for Log Analysis #271)
  • Python 3.13 compatibility (Python 3.13 Compatibility Testing #272)
• Main use-case for me: clean CLI UX, dev/ML workstation setup, and script-friendly commands.

---

4) Would you contribute upstream?

• Yes.
  • I’ve already contributed [MVP] Pre-built package stacks (ML, Web, DevOps) #252 and some small fixes.
  • Actively working on cortex doctor: System Health Check #245 and planning PRs for Stdin Piping Support for Log Analysis #271 and Python 3.13 Compatibility Testing #272.
  • I’m okay writing tests and docs along with code (did this for PR [MVP] Pre-built package stacks (ML, Web, DevOps) #252).

---

Quick suggestions (priority order)

1. Fix API key UX in the wizard (if not already fixed).
2. Polish terminal output (small change, big UX win).
3. Keep code thin and move logic into modules - makes the code easier to read and understand.

---

Happy to help.
Thanks!

[Sahilbhatane]

I didn't check all issues posted but just from use and how most AI related companies like openAI uses AI, I guess we can implement it same way.
Currently we are calling LLM providers through API's, this API are good but here is the problem-

• users usually like no manual work, for current MVP users need to manually add there API to use cortex.

My suggestion -

• create our own LLM ( as I am a LLM engineer i give a quick guide)
• we can fine tune a LLM model relatively large (1-2 B in size) for only LINUX commands.
• we can then deploy this LLM as a private cortex-LLM which will then integrate with or existing project
• this will solve any manual work and we have full control over the AI
• plus we have our own AI

I can implement all this all I need is resources, LLM fine-tuning need large computational power and GPUs as well, i just need this resource and I can make our own LLM which may/can be better than existing API based calling LLM(openAI, anthropic and kimi-k2)