seePickle

Your personal achievement visualization tool.

Comment

Inspiration

🌱 As humans, our ability to solve problems calls for a plan that grows in complexity as the task does. Most importantly, a plan must be envisioned for it to become scalable and successful.

🤖 Now what if this foundational step were to be automated? By harnessing the power of AI, seePickle gets your ideas up and running by generating a base on which your plans can scale.

What it does

🥝 "Node" based plan generation

  • Prompt seePickle the objective of your plan, and it will create a plotted approach for accomplishing your goal(s).
  • For the "Developer" and "Academia" plan types, seePickle can fetch online resources for you to follow along.

👥 Plan assistance

  • For each task, seePickle's assistant can help you with things such as managing costs to generating additional subtasks.

How we built it

Next JSTypeScriptFlaskPythonMongoDB

Secure access and customer identity with Auth0.

This application leverages Cohere's Chat API with Retrieval-Augmented Generation to create a plan for each user prompt.

React Flow was used to build the UI of the plan layout. The Quill.JS wysiwyg editor was implemented for task editing.

Challenges we ran into

Data organization

  • seePickle is a productivity visualization tool where subtasks and miscellaneous data are stored within main tasks, which are subsequently stored within the intended plan.
  • We decided to implement MongoDB's binary json document format for hierarchical data storage, and optimization of CRUD operations.
  • We designed the schema of our database to facilitate efficient exporting of plan data to third party integrations (Notion and Gmail coming soon).

Accomplishments that we're proud of

⚗️ Leveraging Chat with RAG to generate a node-based plan structure in our UI

🚀 Creating an interactive AI-powered Software as a Service

What we learned

Front End: Reactive state management, UI design. Protection of sensitive routes with user authentication.

Back End: CRUD endpoints. Creating string-searching algorithms to parse the Chat response into a python data structure.

What's next for seePickle

⚡ Integration with third party applications (Notion and Gmail)

🧠 Implementation of divide and conquer algorithms for advanced plan generation

Share this project:

Updates