Beamup Social

• 

Inspiration

As someone who works with public companies, I kept seeing the same pain point again and again: Every time a company publishes a press release, someone on the team has to manually convert it into platform-specific social media content, send it for approval, schedule it, and track performance — often across disconnected tools. On top of that, these companies, rarely have dedicated social network manager. Nor they have times to dedicate time to engage their investors, but on the other hand they must, as they profit from them.

I thought: They have knowledge about their business, but they don't know how to engage their investors. This app could solve this issue for them. What if a platform could automate that entire workflow — using AI and a smart, collaborative UI — built specifically for the needs of investor relations and corporate comms teams?

That question inspired Beamup Social: A collaborative, AI-powered content platform that turns press releases into ready-to-go social posts with built-in approval, publishing, and analytics — all inside a secure, multi-company workspace.

What it does

Beamup Social is a web-based platform designed for small to mid-sized teams at public companies. It helps them:

🔄 Convert press releases into AI-generated social media posts instantly

📝 Collaborate through in-app and email-based approval workflows

📤 Publish posts to LinkedIn, Twitter/X, and Facebook, with scheduling and retry handling

📊 Track performance with real-time analytics and URL click tracking

🧠 Chat with an AI assistant trained on company data for content ideas and tone suggestions

🏢 Manage multiple companies and workspaces under one user account

📂 Upload and parse PDF/DOCX files for content generation

📧 Get automatic notifications via Mailgun for invites, approvals, changes, and security events

🔐 Maintain security and compliance, including GDPR support and token usage tracking

It’s everything a public-facing content team needs — streamlined into one clean, fast, and smart interface.

How I built it

I used Bolt.new for fast prototyping, connecting Supabase for my backend and Mailgun for email communication. Here's a breakdown:

🌐 Frontend Built with Bolt’s visual builder

Responsive UI with light/dark mode

Multi-company workspace structure with a company switcher

AI post generator and chat assistant using OpenAI

Rich editor with platform previews, file upload, media picker

🔙 Backend (Supabase) Auth with email/password and role-based access (Global Admin, Team Lead, Member)

Tables for users, companies, posts, media, short links, prompts, analytics, and token usage

Soft-deletes and export logic for GDPR/CCPA compliance

📧 Email Notifications (Mailgun) Invite acceptance

Approval actions (approve/reject/request changes)

Password resets and security alerts

Email + in-app notifications with deep links

🧠 AI Features "Chat with AI Assistant" trained on company context

Document parser for PDF/DOCX to auto-extract content

AI-generated social copy with tone control

Company-based token usage metering

📊 Publishing & Analytics OAuth integration with LinkedIn, Twitter/X, and Facebook (mocked where necessary)

Custom URL shortener with click tracking

Post metrics like impressions and engagement (mock + real blend)

Challenges I ran into

Managing Multi-Company Logic It was tricky designing a clean UX where users could belong to multiple companies, switch workspaces, and maintain separate data scopes for each. Bolt.new made this easier with dynamic routing and conditional components.

✉️ Email + In-App Notification Sync I wanted actions (like post approvals) to work from both the app and email, which meant building secure links, embedding feedback UIs, and syncing approval states — all without compromising UX or security.

🧾 Document Parsing Parsing and cleaning PDF/DOCX press releases proved challenging. I had to combine OpenAI summarization with formatting logic to get clean AI-ready content.

🧠 Token Tracking I wanted to track AI usage per company across both the post generator and AI assistant. This required building a flexible token metering system with monthly caps and warnings.

Accomplishments that I'm proud of

What I learned

How to structure multi-tenant systems with role-based access using Supabase

How to design collaborative UIs that feel clean and intuitive, even with multiple roles and approval flows

The importance of pairing AI automation with human-in-the-loop controls

How to scope a feature-rich product in a limited timeframe by focusing on high-impact flows first (editor > approval > analytics)

What's next for Beamup Social

Post-hackathon, I’d love to: Complete project fully, from some dummy to real data,

Enable AI training based on actual past posts and brand tone,

Expand platform support to include YouTube Shorts and TikTok,

Add smart “best time to post” recommendations,

Offer a public-facing analytics portal for published posts.

Built With

• bolt.new
• react
• supabase