AutoPost.AI

AutoPost.AI – Your AI-powered assistant for generating and sharing viral social media posts in seconds.

Comment

About the Project

What Inspired It

Social media managers, content creators, and small business owners often struggle with writer’s block, platform fatigue, and managing multiple accounts. We wanted to build something that empowers them to quickly generate high-quality content and post it seamlessly. The idea for AutoPost.AI was born from this need—to serve as a smart, AI-driven assistant that makes social media content creation faster and easier.

What We Built

AutoPost.AI is a full-stack web application that allows users to:

  • Enter a topic or campaign idea
  • Generate social media posts using Anthropic Claude 3 Sonnet via Amazon Bedrock
  • Connect their Twitter account via OAuth
  • Copy or directly post content to Twitter (with plans to support LinkedIn/Discord)
  • Toggle between light and dark modes
  • Manage connection status via real-time token and cookie storage

How We Built It

  • Frontend: React + Vite + TailwindCSS + MUI for the UI, deployed on Vercel
  • Backend: Python-based AWS Lambda functions for all logic (OAuth, post generation, Twitter posting), deployed using AWS SAM
  • AI Integration: Used Amazon Bedrock to call Claude 3 Sonnet for high-quality, human-like text generation
  • API Gateway to expose Lambda functions as secure REST endpoints
  • IAM Roles to handle permissioned access
  • DynamoDB (planned) for storing session/token data securely
  • OAuth 1.0a for Twitter authentication using requests-oauthlib

What We Learned

  • How to securely integrate Bedrock in a Lambda environment
  • Handling OAuth flows within a serverless architecture
  • Optimizing serverless deployments using AWS SAM
  • Dealing with CORS, environment variables, and credentials securely
  • Implementing real-time UI feedback and transitions with token-based state

Challenges We Faced

  • Twitter's 1.0a OAuth and token callback handling was unexpectedly tricky due to redirect and callback URL validation.
  • Managing local development vs. deployed testing required multiple fixes for CORS, HTTPS issues, and credentials handling.
  • We encountered issues with unverified OAuth apps (especially LinkedIn) and had to adjust expectations and UI flows accordingly.
  • Invoking Bedrock’s Claude 3 via Lambda required careful JSON formatting and model-specific constraints (like prompt roles and max tokens).
  • Faced issues with the auto-post tweets into twitter since this advanced access was available in upgraded twitter plan.
  • Faced issues while deploying the app, iam permissions, stack issues, bedrock connectivity issues etc.

What’s Next for AutoPost.AI

1. Multi-Platform Auto-Posting Support

  • Current: Twitter supported via OAuth 1.0a.
  • Next: Add full OAuth and posting support for LinkedIn and Discord.
  • Goal: Allow users to connect multiple platforms and choose where to post with one click.

2. Scheduled Posting (Auto Queue)

  • Let users schedule posts by selecting a date/time.
  • Use Amazon EventBridge + AWS Lambda to trigger posting tasks on schedule.

3. Content Personalization with Templates

  • Offer style presets like “Professional”, “Casual”, “Witty”, or “Minimal”.
  • Users select a tone, and prompts are adjusted automatically before generation.

4. Post Preview and Rich Media Integration

  • Add preview UI with platform-specific formatting.
  • Enable AI-generated hashtags, emojis, and optionally image generation via Amazon Titan or DALL·E.

5. Post History and Dashboard

  • Build a personal dashboard (with user sessions) to:

    • View previously generated posts
    • Track which posts were published where
    • Reuse or repost content

  • Store data using DynamoDB with user ID as a partition key.

6. Analytics and Engagement Estimation

  • Use AI to estimate reach, engagement, or recommend best posting time.
  • Potential integration with platform analytics (via API or mock feedback).

7. Mobile-Optimized UI

  • Build a fully responsive version for mobile users.
  • Consider creating a companion PWA (Progressive Web App).

8. Team Collaboration Mode

  • Allow users to invite team members to a shared workspace.
  • Share prompts, approve posts, or assign platforms (like Buffer, but smarter).

Built With

Share this project:

Updates