Hooser

Automatically generate detailed user stories as per your custom templates and push them as issues to JIRA.

Comment

Inspiration

During my stint as a product manager, one of the key tasks I performed while building products was preparing detailed user stories. This was something that was equally dreaded by everyone in the team as documenting detailed user stories consumed a lot of our time. We would rather use that time engaging with users and customers, or diving deep into the technical designs with architects or even be involved in the sales and marketing processes. I was looking for a way to automate some of these repetitive tasks and that's how the idea came about.

What it does

Hooser helps product managers drastically reduce the time they take to document user stories and create issues in Jira-like ticketing systems. You simply give a very brief description of the feature that you are looking to build, then pick a pre defined template that has guidelines for writing user stories, including what language style to use, formatting to use and the sections to include and Hooser will do the rest for you. Hooser will generate a detailed user story, based on your description of the feature and the formatting guidelines. If you're happy with the result, you can then create a JIRA issue with the user story at the click of a button.

How we built it

The multi-page app was written in Python and built using Streamlit. It is hosted in Streamlit Community Cloud. Snowflake Arctic was the foundational LLM used for text generation. The model was accessed using Replicate's APIs.

Challenges we ran into

Faced some issues with getting Snowflake Cortex APIs to work locally (installation of dependant packages were failing). This was due to the latest version of Python not being supported. Solved it by installing an older version initially, then moved to using Replicate APIs to access the LLM.

Accomplishments that we're proud of

New to the world of Python and LLMs. Happy to see a working app built in such a short span.

What we learned

I am new to Python programming. After playing around with some scripts and getting a bit familiar with Python, I was looking for a framework for building apps in Python. I was trying to decide between Flask and Django - that's when I came across Streamlit. Streamlit has a much simpler learning curve and produces higher impact results faster. Along with the Community Cloud, this is a great way to build and deploy apps fast and for free.

What's next for Hooser

From a feature perspective, the roadmap for Hooser has many exciting items.

  1. Improve template management.
  2. Integrate with Linear
  3. Incorporate authentication and user accounts.
  4. Incorporate RAG to ground user stories on custom user research and data insights.
  5. Tagging and storage mechanism for user research and insights dataset.
  6. Workflows to create new products, offerings and batch generation of user stories.

Anybody wants to join forces? :)

Built With

  • arctic
  • python
  • replicate
  • snowflake
  • streamlit
Share this project:

Updates