DevSparkAI Social Hub

Planning and producing consistent, platform-specific social content takes significant time for small teams and businesses.

I wanted to build a product where AI is part of the core workflow — planning, writing and adapting content — not a bolt-on chat box.

Users provide a topic and timeframe, and the system generates a structured content plan, then platform-specific variants that can be edited, scheduled and published.

Provider abstraction layer

All AI calls go through one provider layer, so OpenAI, Gemini and other compatible providers are interchangeable without changing product logic.

Bring your own key — encrypted

Users can supply their own provider API keys. Keys are encrypted before storage and secrets are kept in environment configuration.

OAuth2 PKCE for X/Twitter

Publishing to X uses the OAuth2 PKCE flow, so the app never handles user passwords.

• AI content generation & planning
• Brand profiles
• Platform-specific content variants
• Post scheduling & publishing
• Analytics
• Multiple AI providers (OpenAI, Gemini, others)
• Bring-your-own API keys
• X/Twitter OAuth2 PKCE
• Encrypted API credentials
• JWT-protected API routes

• Integrating LLMs into a full product workflow
• Secure credential handling and OAuth flows
• Designing for multiple AI providers from the start

This project demonstrates how I integrate AI capabilities into a complete SaaS product rather than building AI as an isolated prototype.