PostPilot AI

AI agent that runs social media for small businesses — generates posts, captions, and content calendars using Gemini. Zero human effort required.

Comment

Inspiration

Small businesses spend 5-10 hours per week on social media content. Most cannot afford a social media manager ($3,000-5,000/month). I built PostPilot AI to give every small business owner an AI agent that generates a full week of content in 30 seconds — no copywriter, no agency needed.

What it does

PostPilot AI is an AI agent powered by Google Gemini. A business owner enters their business name, type, target audience, and goal. The Gemini agent runs 4 sequential steps:

  1. Analyzes business context and builds a content strategy
  2. Generates 5 platform-specific posts (LinkedIn, Instagram, Twitter/X, Facebook)
  3. Writes optimized hashtags for each platform
  4. Builds a complete 7-day content calendar

Every step is logged with timestamps and token counts in a live agent execution log visible in the UI.

How we built it

  • Frontend: Next.js 14 + TypeScript + Tailwind CSS
  • AI: Google Gemini API (Gemini 3.5 Flash Lite) via AI Studio
  • Database: Firebase Firestore (stores all agent runs as production logs)
  • Auth: Firebase Authentication
  • Hosting: Vercel
  • Google Cloud: Firebase (satisfies Google Cloud platform requirement)

Challenges we ran into

Finding the right Gemini model version that worked reliably on the free tier required iteration through multiple versions. The second challenge was making Gemini output structured JSON reliably for posts and calendar data — solved with strict prompt engineering and a fallback JSON parser.

Accomplishments that we're proud of

The live agent execution log is something I'm particularly proud of. Every Gemini API call is logged with its timestamp, action, token count, and output — making AI decision-making fully transparent to users and judges alike. The app generates real, usable social media content in under 30 seconds.

What we learned

Building AI-native means every core user action goes through the AI — not just one feature. The agent log pattern separates a product with AI bolted on from a genuinely AI-native product. Prompt engineering for structured JSON output is harder than it looks but critical for reliable production use.

What's next for PostPilot AI

  • Stripe payment integration for Pro ($19/month) and Business ($49/month) tiers
  • Direct social media publishing via LinkedIn and Instagram APIs
  • Brand voice training — upload past content and Gemini learns your tone
  • Scheduled posting with AI-optimized timing recommendations
  • Multi-language support for non-English businesses

Built With

Share this project:

Updates

posted an update —

PostPilot AI is now live.

We built PostPilot AI to help small businesses turn their marketing goals into ready-to-use social media content using Gemini.

Key features:

  • AI-generated platform-specific social posts
  • AI-powered content strategy
  • 7-day content calendar
  • Agent execution history and logs
  • Simple workflow designed for small businesses

Our approach is to use Gemini as an AI marketing operator rather than simply as a chatbot. PostPilot analyzes business context, develops a content strategy, generates platform-specific content, and organizes the campaign into a weekly plan.

Built for the Build with Gemini XPRIZE Hackathon.

We are excited to share the product and continue improving it based on user feedback.

Log in or sign up for Devpost to join the conversation.

Submission history