GrantBeacon

• 

  Hero

• 

  Mission

• 

  Login screen

• 

  Feed Page

• 

  Shortlist Page

• 

  Topical Filters- Senior Care Sector

• 

  AI Grant Summary

• 

  Edit Profile Page

• 

  Admin Dashboard

• 

  Admin Panel

• 

  Admin Panel 2

About the Project

“I’m a management professional who’s never written a line of code. With Bolt’s AI coding agent I built a production-ready platform in weeks, proving that AI truly empowers non-technical creators.”

Inspiration

• Mission need. 35 k senior-living (assisted living) communities in the U.S. care for ≈ 1 million older adults, yet 25 % (mostly nonprofits & faith-based) struggle to secure the millions in grants released each year because the data is scattered and staff time is scarce. Staff spend hours hunting grants, often days late or unaware of niche programs. Result: lost funding, stalled innovation, and widening equity gaps between well-resourced chains and small nonprofits.
• Personal lens. Years of consulting inside senior care showed me just how many Senior Care programs die on the vine for lack of funding.
• AI empowerment. Seeing Bolt’s vibe coding model let a non‑technical founder turn an idea into a live product made me jump in head‑first.

What it does

GrantBeacon unifies grant discovery, AI summarization, and senior‑care insight into a single, seamless platform:

1. ETL → Supabase
   Bolt Edge Functions fetch, de‑duplicate, and enrich fresh data from Grants.gov & SAM.gov regularly, tagging each record with senior‑care keywords.

2. AI‑Powered Summaries (OpenAI GPT‑4o)
   A callable function condenses grant information into 200‑word, plain‑English digests covering scope, eligibility, award amounts, and contact info—so administrators scan insight, not PDFs. Ability to look at actual grant using on page URL

3. Industry‑Specific Filters
   Agency type , program focus (Dementia, Tele‑health, Workforce), cost sharing toggles, and award range sliders zero‑in on the grants that truly fit senior‑living providers.

4. One‑Click Shortlisting
   Users save promising grants to revisit in future

5. Admin Dashboard
   External Grant uploads, source health, debug tools

Feature Overview:

How we built it

Our Journey

1. Ideation and Brainstorming
   With OpenAI’s o3 model, I was able to rigorously stress-test my ideas across multiple dimensions. From the start, my goal was to build something meaningful—something that supports the senior living communities that care for our elders and serve as the backbone of our society. I wanted to channel my years of experience in the Senior Care industry into a solution that could move from concept to a real, impactful product. After an intensive brainstorming process and narrowing down several possibilities, GrantBeacon was born. To ground the idea in real-world needs, I also held focused conversations with CXOs in the senior living space (through my network), which affirmed both the critical need for a tool like GrantBeacon and the gap it is designed to fill.

2. From PRD to initial Product
   Bolt’s AI Coding Agent transformed my rough PRD—mindfully crafted with OpenAI’s o3—into a functional v1 product. I broke the PRD down into broad, structured steps, translating them into prompts that could be fed directly into Bolt’s agent for seamless execution.

3. Data Ingestion Pipeline
   As a non-technical founder, this was a significant undertaking. I began with in-depth research into the grant sourcing landscape and identified two key sources—Grants.gov and SAM.gov—for the v1 product. Using Bolt, I set up edge functions with Supabase, making the product highly scalable and efficient by integrating with these APIs. This marked a major milestone in the journey of GrantBeacon. To ensure relevance, I implemented industry-specific screening and filtering to surface only the most applicable grant opportunities. Looking ahead, the roadmap includes expanding to additional data sources to further strengthen the platform.

4. AI Summaries
   Watching GPT-4o distill complex grant information into a concise 200-word digest convinced me that AI-powered summaries were the killer feature. This became the second major milestone in the journey—one that adds immense value for the end users who need clarity without the clutter. With Bolt, I was able to set up edge functions in Supabase to automatically generate these summaries, bringing this vision to life.

5. Ship, share, and iterate
   One-click deploy pushed GrantBeacon to grant-beacon.org under a custom domain; research and user feedback now guides the roadmap toward Invoke-AI proposal drafting and broader data coverage.

Challenges we ran into

1. Non‑technical founder hurdles.
   My journey started with understanding how to setup github, supabase accounts and learning for the first time about Edge functions. It meant plenty of late‑night learning and debugging sessions (Pro tip: keep ChatGPT open in another tab!)

2. API errors
   Setting up the Data ingestion APIs was time consuming; mapping, retry logic, and occasional errors kept the Edge Function console lively

3. Credit exhaustion.
   This was a major hurdle—one that almost made me consider giving up. I had exhausted my initial Bolt credits while setting up the platform and troubleshooting early errors. As the context window grew, each query began consuming a large number of tokens, making iteration costly. To push through, I had to get creative. I used the GitHub portal to convert my codebase into text files, which I then fed into ChatGPT/OpenAI to keep building and iterating efficiently. It became a scrappy but effective way to move forward without stalling the project

4. Link accuracy.
   Early NOFO URLs from search2 pointed to generic pages; digging into the docs and switching to opportunityId endpoints fixed the judge‑click issue

Accomplishments that we're proud of

• Zero-to-Launch in Weeks — Went from a blank Bolt workspace to grant-beacon.org live on a custom domain, proving that a non-coder can ship production software with AI. 175 + grants live.

• AI Summaries That Save Hours — Integrated GPT-4o streaming to condense dense grant information into 200-word digests, turning multi-page government jargon into quick, actionable insights for busy administrators

• Mission That Extends Beyond the Hackathon — The early traction fuels our roadmap: Invoke-AI proposal drafts, broader data coverage, and win-probability scoring aimed at unlocking millions for senior-care innovators

What we learned

• AI lowers the barrier — Prompt‑driven coding let a business mind ship full‑stack features without traditional engineering resources.

• Data integrity fuels every decision — A fault-tolerant, ETL pipeline keeps the grant feed current and de-duplicated, so providers act with confidence instead of second-guessing spreadsheets.

• Purpose-built sparks action — The deeper mission of empowering senior care providers kept me motivated and helped me push through every challenge along the way.

What's next for GrantBeacon

• Invoke AI — Using AI to auto‑draft grant proposals by reviewing grant information and org profile. This will help create grant applications automatically.

• Broader ingestion — state portals (CA, NY), foundation RSS, corporate CSR uploads

• Win‑probability scoring & Slack / email alerts < 1 hr from publish

• Mobile PWA for on‑the‑go grant scouting

📊 Impact to Date

🏆 Challenge/Bonus Participation

Built With

• bolt
• chatgpt
• entri
• github
• grants.gov
• javascript
• lucide-react
• markdown
• netlify
• openai
• postgresql
• react
• sam.gov
• sql
• supabase
• tailwindcss
• typescript
• vite