We've all been there. You start a new project and receive a folder filled with 50-page PDFs, complex diagrams, and scattered Markdown files. You just want to find one specific detail, like why a deployment might be failing or where an obscure config setting lives. But you're stuck scrolling through thousands of lines of text, hoping the answer is in there somewhere. It is incredibly frustrating to have to stop mid-flow just to hunt for a requirement buried in a three-year-old document. That time spent digging is time you aren't actually building, and honestly, it is a massive drain on the creative energy you need to ship great code.
Generic AI chat tools try to bridge this gap, but they often lack the one thing developers need most: context. A general large language model (LLM) doesn't know your team's internal nuances and quirks. To solve this, I am excited to introduce personal AI notebooks (now in developer preview) within Red Hat Developer Lightspeed on Red Hat Developer Hub.
What are personal AI notebooks?
You can think of this notebook as a dedicated knowledge base for a specific task. Rather than using an AI that knows the entire internet but nothing about your internal architecture, you can create a workspace, upload your specific project documents, and engage in a conversation grounded strictly in that data.
Retrieval-augmented generation (RAG) powers the personal AI notebooks. When you ask a question, the system looks through your uploaded files, finds the relevant passages, and connects the dots for you, giving an answer that actually makes sense in your specific context (rather than just making lucky guesses).
Why does this matter for developers?
This feature helps developers in three main areas:
- Source transparency: Every claim the AI makes comes with a sources chip. One click shows you the exact file and snippet from where the AI pulled. This means you're not just taking its word for it because it provides the evidence instantly.
- Data isolation: Your onboarding notebook does not talk to your security audit notebook. This prevents context bleeding and ensures your queries remain relevant to the task at hand.
- Instant context: New team members can stop waiting for a senior developer to be free and instead ask the notebook, "What are the core architectural principles of this service?"
Under the hood: How it actually works
If you're an architect or platform engineer, you'll be glad to hear that the pipeline behind these notebooks is tuned to handle heavy document processing without dragging down performance.
Notebooks in Developer Hub is built on top of Developer Lightspeed and uses its resources and configuration. The backend utilizes a vector database to store document embeddings, the same way Developer Lightspeed chatbot vectorizes its Developer Hub documentation. When you upload a file, the system converts it into mathematical vectors that the AI can read efficiently. To get the best balance of speed and reasoning, I recommend using GPT 4.1 model or higher.
During this developer preview, we have also implemented resource guardrails to keep the experience snappy:
- Maximum file size: 25 MB per upload
- Privacy: Each notebook is private to the individual user
- Hallucinations: AI responses are based strictly on your documents
Set it up for administrators
If you are a platform engineer enabling this for your team, it is a simple configuration change in your Developer Lightspeed Helm values.yaml settings.
lightspeed:
notebooks:
enabled: true
queryDefaults:
model: gpt-4.1
provider_id: openaiWe designed the system to be resilient. If you make a typo in the model name during setup, Developer Lightspeed will surface a helpful error message in the logs and user interface to help you troubleshoot.
Clean up your experience
I focused on making the new notebooks tab feel like a seamless part of your workspace rather than just another feature added to the menu. The new dashboard organizes your projects into simple cards, so you can see exactly what's inside and when you last touched a topic without digging through menus.
I also overhauled the prompt bar to make swapping models and attaching files feel like a natural part of the conversation rather than a chore. The goal is to make the transition from a general chat to a deep dive into your own project docs feel seamless, keeping the focus entirely on what you need.
Give it a spin
Documentation should not be a roadblock. We built personal AI notebooks on the idea of turning those dense, static files into an actual conversation. At the end of the day, it's about getting the answers you need quickly so you can spend your energy on what you actually enjoy, building great software.
Ready to dive deeper? Explore our learning paths for Developer Hub.
