InvestCast

InvestCast helps you truly understand the companies you want to invest in with podcast-style summaries and deep dives, based only on official public records. Think Spotify meets investing.

Comment

Inspiration

My inspiration for InvestCast emerged directly from my tenure as a Senior Product Manager within the Canadian investment brokerage sector. During that time, I keenly observed a significant challenge: many investors, whether novices or seasoned professionals, often lacked a true understanding of the companies they were investing in. The primary culprit? The sheer difficulty and time commitment involved in conducting deep dives or even gaining more than a superficial understanding of a company's operations.

InvestCast's core mission is to dismantle this barrier. I envisioned an application that would provide an effortless way for individuals to quickly explore available companies, grasp their fundamental business models, assess their potential fit for personal portfolios, and for premium subscribers, access in-depth analyses to inform their decisions. My personal investment journey further solidified this vision; my portfolio performance significantly improved once I dedicated time to a deeper understanding of my investments. I was driven to replicate this benefit for others, but in a far more accessible format, eliminating the drudgery of manual research.

What it does

InvestCast is an innovative financial podcast platform designed to empower investors and curious minds with accessible, on-demand audio insights into companies across various sectors. The application currently offers the following core functionalities:

  • Sector-based Company Discovery: A dedicated "Sector Page" allows users to browse companies by industry, providing a streamlined way to discover new investment opportunities or research specific sectors. Companies can be searched by name or ticker and sorted by key metrics like stock price and market capitalization. Each company listing provides key financial data such as current price, price change, price change percentage, and market cap.
  • Deep-Dive Company Podcasts: For each company, InvestCast presents a collection of relevant podcast episodes. These include concise "Company Snapshots" and more in-depth "An Investor's Guide" analyses. These podcasts aim to deliver crucial financial and strategic information in an audio format. "An Investor's Guide" episodes feature expandable descriptions for comprehensive reading.
  • Personalized Library & Watchlist: Users can curate a personal "Library" of their favorite podcast episodes and maintain a "Watchlist" of companies they are actively tracking. This allows for quick access to relevant content and streamlined monitoring of companies of interest.
  • Seamless Audio Playback: A persistent audio player is available at the bottom of the screen, displaying the currently playing podcast's title, company name, and ticker. It provides standard playback controls including play/pause, skip back/forward by 10 seconds, and volume adjustment. The player also features a scrubbable progress bar.
  • User Interaction with Content: From both the company podcast page and the audio player, users can add podcasts to their library (favorite them) and add/remove the associated company from their watchlist.
  • Feedback Integration: The AudioPlayer includes a "Report Issue" button, facilitating direct user feedback on content or technical aspects.

How we built it

InvestCast was built with a keen focus on leveraging AI for content generation and Bolt.new for rapid frontend development and integration.

  1. AI-Powered Content Generation: The core of InvestCast's unique content lies in its AI-driven research and scripting. We employ various AI models, including Google Gemini, to conduct deep dives into financial reports, news articles, and market data. The insights gathered are then synthesized by other specialized AI models to generate structured, engaging scripts for "An Investor's Guide" and concise summaries for "Company Snapshot" podcast episodes. For the episodes, we utilized Google Gemini for high-quality, natural-sounding voiceovers.
  2. Automation with n8n: The entire content generation pipeline is orchestrated by an intricate workflow built using n8n. This automation handles everything from initiating AI research and managing script generation to triggering text-to-speech conversion and uploading the final audio and metadata to our Supabase database.
  3. Supabase Backend: All critical data, including user profiles, podcast metadata, company details, and user-specific library and watchlist information, is securely stored and managed within a Supabase database. This provides a scalable and robust foundation for the application.
  4. Bolt.new Frontend Development: The entire user interface of InvestCast, encompassing the Sector Page, Company Podcasts view, User Library, Watchlist, and the interactive Audio Player, was primarily developed using Bolt.new. Bolt.new’s ability to take initial product requirement documentation and even AI-generated UX design documentation as a starting prompt significantly accelerated the initial build, providing about 80% of the foundational structure within the first few hours. Its iterative development capabilities and seamless integration with our Supabase backend were instrumental in bringing the project to its current state.

Challenges we ran into

Putting together a project of this nature, while seemingly straightforward in concept, presented several significant challenges:

  • Perfecting AI Prompts and Model Selection: A major hurdle was perfecting the prompts required to conduct comprehensive research, accurately synthesize information, and generate engaging scripts and summaries. It was particularly challenging to identify the optimal AI models for specific research and content creation tasks, each requiring nuanced prompting strategies.
  • Complex N8n Automation Flow: While n8n initially appeared simple, building a robust and complex automation flow capable of powering a sophisticated content generation engine proved far more time-consuming than anticipated. The intricacies of creating a seamless, end-to-end pipeline demanded meticulous attention to detail and extensive debugging.
  • Content Quality and Accuracy: A critical challenge was ensuring that the AI-generated content was not only engaging but also factually accurate and unbiased, especially given the sensitive nature of financial information. Implementing robust validation steps within our pipeline to mitigate AI hallucinations was a continuous effort.
  • Balancing Automation with Oversight: Achieving the desired level of automation while maintaining essential human oversight for quality control and critical review of AI-generated financial content presented a delicate balance.

Accomplishments that we're proud of

We are incredibly proud of several key accomplishments with InvestCast:

  • Successful AI-Powered Content Generation: We successfully implemented a complex AI pipeline that can research, script, and generate audio content for financial podcasts, significantly streamlining content creation.
  • Intuitive User Experience: We've built a clean, intuitive, and accessible user interface that makes exploring complex financial information approachable for a wide audience. The feedback modal integrated into the audio player provides a direct channel for continuous improvement.
  • Rapid Development with Bolt.new: The speed and efficiency with which we were able to develop the frontend using Bolt.new were truly remarkable. Starting with AI-generated product and UX documentation, Bolt.new enabled us to establish 80% of the foundation very quickly and iterate effectively.
  • Comprehensive Financial Data Display: The integration of real-time (or near real-time) financial metrics for companies directly within the app provides immediate value to users.
  • Scalable Architecture: By leveraging Supabase and designing our content generation pipeline with automation in mind, we've laid the groundwork for a scalable platform capable of handling a growing user base and extensive content library.

What we learned

This hackathon provided invaluable learning experiences:

  • The Depth of Prompt Engineering: I learned that prompt engineering is a deeply nuanced skill, requiring extensive experimentation and iteration to achieve desired outcomes from AI models for specific tasks like financial research and script generation.
  • Complexity of Automation Flows: The true complexity of building sophisticated automation workflows, even with tools like n8n, became apparent. Nuances in data handling, error management, and sequential processing required a much deeper understanding than initially perceived.
  • The Power of Bolt.new: While I had prior exposure to Bolt.new, this project solidified my understanding of its immense power, especially in rapidly translating design and product requirements into functional code. Its ability to suggest and create pages/interactions beyond initial scope was particularly enlightening.
  • Interdisciplinary Integration: Successfully combining AI, automation, frontend development, and database management into a cohesive, functional product highlighted the importance of interdisciplinary understanding and seamless integration.

What's next for InvestCast

InvestCast has an ambitious roadmap ahead:

  • Advanced AI Features: We plan to integrate even more sophisticated AI models for real-time sentiment analysis of news and social media, providing users with instant insights into public perception. We also aim to develop a conversational AI interface where users can ask specific financial questions and receive audio-based answers.
  • Personalized Content Recommendations: Implementing an AI-driven recommendation engine that learns from user behavior to suggest highly relevant companies and podcast episodes will significantly enhance content discovery and engagement.
  • Expanded Content Library: We will continuously expand our database of companies and generate more "An Investor's Guide" and "Company Snapshot" episodes, covering a broader range of sectors and geographies.
  • User Engagement Features: Introducing features like listening progress tracking ("Recently Played" filter in the Library) and the ability to download episodes for offline listening will improve the overall user experience.
  • Monetization Enhancements: Further developing the premium subscription model with additional exclusive content and features will be a key focus.
  • Community and Social Features: Exploring options for users to share insights, discuss companies, and engage with financial content in a community setting.

Built With

Share this project:

Updates

Submission history