I got tired of wiring up auth, database migrations, file uploads, and WebSocket channels for every new project. Supabase's self-host stack is a dozen containers. PocketBase is SQLite. I wanted PostgreSQL, supabase-js compatibility, and a deploy button that works.

So I built Sinopebase — a single Bun service that speaks the supabase-js API. Auth, database, storage, realtime, and AI in one process. Deploy it on Railway, connect PostgreSQL and a bucket, and you have a working backend before your coffee gets cold.

What you get:

• Auth — email/password, OAuth (Google, GitHub, Discord, Apple, Microsoft), enterprise OIDC
• Database — PostgreSQL with row-level security, migrations, realtime subscriptions
• Storage — S3-compatible (Railway buckets just work)
• Realtime — Phoenix Channels with presence (track, untrack, heartbeat)
• AI — Mastra agents, RAG, streaming — built in, not bolted on
• Admin UI — table editor, user management, storage browser, API docs, metrics, logs
• SDK — drop-in compatible with supabase-js. One import swap.

// was: import { createClient } from '@supabase/supabase-js'
import { createClient } from 'sinopebase-js'

const sb = createClient('https://your-service.railway.app', 'your-anon-key')
const { data } = await sb.from('todos').select('*')

The pitch: If you know supabase-js and you want a backend you own — without a dozen containers, without per-seat pricing, without vendor lock-in — this is for you.

Pre-1.0, honestly. I'm a solo founder. This powers my own product. It's production-like but I haven't done the supply chain attestation or external audit yet. Everything works. If something doesn't, open an issue and I'll fix it.

Deploy: [Railway template link]
npm: npm install sinopebase-js
Repo: https://github.com/KalvexDotDev/sinopebase

Questions, roast, and "why didn't you just use X" all welcome.