Abeshare Social

Engineering case study

One post.Ten platforms. Fully automated.

I designed and built a production full-stack platform that composes a post once, generates AI-tailored versions for each network, and publishes or schedules to 10 social platforms at the same time — end to end, solo.

10
Social platforms integrated
4
Containerized services
1 GB
Max media per post
$30K+
Comparable agency build

The problem

Posting the same content to every platform by hand is slow and repetitive, and subscription schedulers are expensive, rented, and limited in what they connect to.

The solution

A self-hosted platform: one compose box, AI variants per network, publish or schedule everywhere at once. Fully owned, no per-seat fees, complete control of data and accounts.

Integrated networks

10 platforms, one workflow

InstagramFacebookYouTubeTikTokLinkedInX (Twitter)ThreadsBlueskySpotifyWhatsApp

What it does

Built for real publishing

AI captions per platform

Write once; a multi-provider AI layer (Groq, DeepSeek, OpenAI/Grok, local Ollama) generates copy tailored to each network with automatic failover.

Schedule or publish instantly

Queue posts for the perfect time or fire immediately. A background worker handles delivery on schedule.

Large media + Google Drive

Upload up to 1 GB or paste a Google Drive link — the server downloads and distributes it to every platform.

Reliable job queue

An async worker with retry logic and per-job status tracking ensures failed posts surface clearly instead of disappearing.

Secure OAuth connections

One-click account linking via OAuth 2.0 and 1.0a, with token storage and automatic refresh for each provider.

App-review compliant

Privacy, Terms, and GDPR/Meta data-deletion flows built to pass Meta and TikTok third-party app review.

How it's built

Decoupled, queue-based architecture

The web tier and the posting worker are independent services communicating through a job queue, so the UI stays responsive and delivery is reliable and retryable.

Web app

Next.js dashboard + API routes

PocketBase

Auth, data, media storage, job records

Worker

Polls the queue, posts to each platform

Cloudflare Tunnel

Secure HTTPS, zero open ports

Tech stack

Tools & technologies

Frontend

  • Next.js 15 (App Router)
  • React 19
  • TypeScript
  • Tailwind CSS v4
  • shadcn/ui

Backend & Data

  • PocketBase
  • REST APIs
  • Job queue
  • File/media storage
  • Async Node.js worker

Integrations

  • OAuth 2.0 / 1.0a
  • Meta Graph API
  • YouTube resumable upload
  • TikTok Content Posting
  • WhatsApp Cloud API

AI

  • Multi-provider LLMs
  • Automatic failover
  • Prompt engineering
  • Per-platform variants

DevOps

  • Docker Compose (4 services)
  • Cloudflare Tunnel
  • Playwright automation
  • Secrets management

Architecture

  • Decoupled web + worker
  • Queue-based posting
  • Token lifecycle
  • Graceful degradation

The value

A one-time build vs. forever fees

Comparable SaaS tools (Buffer, Hootsuite, Later) cost $15–$100+ per user every month. This is an owned asset: it runs for a few dollars a month and the owner keeps full control of their data and accounts.

Comparable agency build$30,000+
Per-integration market value$1,500–$4,000 each
Production hosting~$8–25 / month
SaaS alternative$15–100 / user / month

Skills demonstrated

What this project proves

Full-stack architectureAPI integrationOAuth 2.0 / 1.0aAsync workers & queuesLLM / AI integrationDocker & containerizationTypeScriptReact / Next.jsNode.jsREST API designToken lifecycle managementGDPR compliance

Built by Damilola Abe

Full-Stack Software Engineer · Available for full-stack roles and freelance builds