Loading post…
OktaTanish KumarTutorial
How I Built an Okta Documentation Chatbot in Python
Summary
Okta developer documentation contains information developers need for integrations, but the author observed that it was underused. They built Oktanaut, a Python chatbot in Jupyter Notebook, comparing a LlamaIndex-based prototype trained on Okta documentation with a GPT-3.5-powered version using the OpenAI API. The first favored precision for focused questions, while the second handled a broader range; the final demo used OpenAI API for its larger models and user-friendly experience. For both versions, the author curated questions, answers, and support engineer information, stored training data in Google Drive for Colab, and retained historical interactions so replies could use session context. Panel provided the frontend, and the author chose Jupyter Notebook for simple setup despite noting bugs and glitches.
Context
The project was created to help developers access Okta Developer Documentation through a conversational interface. The supplied walkthrough covers training data, API configuration, Colab setup, the Panel interface, and continued conversations using prior session messages.
Approach / What changed
The author built two Python chatbot versions: a focused LlamaIndex prototype trained on Okta documentation and a broader GPT-3.5 implementation using the OpenAI API. The selected version uses curated documentation and support-engineer data, historical conversation context, Panel widgets, and a Jupyter Notebook deployment run through Google Colab.
Takeaways
The LlamaIndex version produced more accurate responses for focused questions, while the GPT-3.5 version handled a broader range of inquiries.
Training data included curated questions and answers plus information from Okta Developer Support Engineers; historical interactions supplied context within a session.
The Colab setup requires OpenAI, Panel, and LlamaIndex packages, an OpenAI API key, and documentation files placed in the expected Google Drive folder.
FAQ
Why did the author choose the GPT-3.5 version over the LlamaIndex prototype?
The LlamaIndex prototype excelled in accuracy but could not handle diverse questions. The GPT-3.5 version produced better responses across a broader set of questions and was considered more user-friendly for the demo.
How does Oktanaut use conversation history?
The chatbot appends user prompts and assistant responses to a context list during the session. It also stores historical interactions so later responses can use context from earlier exchanges.
What is needed to run Oktanaut in Colab?
Users need a Google account with Colab access, the OpenAI, Panel, and LlamaIndex packages, and an OpenAI API key. They must also place the documentation files in an oktanaut folder in Google Drive or update the notebook path.
Source and freshness
- Source
- developer.okta.com
- Published
- Added to Yomu
Summaries are generated from the original by a language model. Check the original before relying on it.
Keep reading
Related reading
The Good Stuff
Paid picks we use ourselves. Links carry a referral code — a subscription through one earns us a commission at no extra cost to you.
The best engineering reads, once a week
One useful email. Unsubscribe anytime.