Video Streaming Backend
Part of the Backend Developer Roadmap
In short
Nobody downloads a whole movie before pressing play. Streaming backends cut video into small chunks at several qualities, so playback starts fast and adapts to your network.
A backend supporting video upload and streaming — practice for handling large files, streaming responses, and range requests for seekable video playback.
Think of it like this
Hot puris at a wedding
The caterer doesn't dump 50 puris on your plate at once. He serves two at a time, hot, as you eat. If you slow down, he slows down. Video streaming serves small chunks the same way, just in time.
Watch it happen
From upload to playback
1/6The creator uploads a large MP4 straight to object storage with a presigned URL.
The big ideas
What you need to know
Transcode in the background
Converting video is heavy CPU work. It runs in queued workers, never inside an API request.
HLS: chunks plus playlists
Each quality is a list of small segments in an .m3u8 playlist. A master playlist lists all qualities.
Adaptive bitrate
The player measures download speed and picks the best quality for the next chunk, so video rarely stalls.
CDN does the heavy lifting
Segments are static files, perfect for CDN caching. Your servers handle metadata, auth and analytics.
See it in code
Walk through the code
1# Turn one upload into a 720p HLS stream: small .ts chunks plus a playlist2ffmpeg -i upload.mp4 \3 -vf scale=-2:720 -c:v libx264 -preset veryfast -b:v 2800k \4 -c:a aac -b:a 128k \5 -hls_time 6 -hls_playlist_type vod \6 -hls_segment_filename "out/720p_%03d.ts" \7 out/720p.m3u88 9# Repeat for 480p and 360p, then list all three in a master playlist,10# so the player can switch quality as the network changes.1/3Scale to 720p height, encode video with H.264 at about 2.8 Mbps and audio with AAC.
Run this inside a worker that picks up transcode jobs. The output folder is what you upload to storage and serve through a CDN.
Save yourself the pain
Common mistakes
Oops
Serving the original MP4 from your Node.js server.
Do this instead
Every viewer then pulls the full-size file through your server. Transcode to HLS and serve from storage plus a CDN.
Oops
Running FFmpeg inside the upload request.
Do this instead
Transcoding can take minutes. Queue it, return processing status, and notify when it's ready.
Your mission
Milestone 1: play your own HLS
Quick check
Why are videos cut into small segments at several qualities?
Go deeper
- 1Support HTTP range requests so users can seek within a video without downloading it fully.
- 2Process uploaded videos asynchronously, e.g. generating thumbnails, via a job queue.
- 3Store video files in cloud storage and serve them through a CDN.
More projects at this stop
URL Shortener
Blogging Platform API
E-commerce Backend
Job Queue System
Rate Limiter Service
Real-time Chat Backend (WebSocket)
Authentication Microservice
Payment Integration Backend (Stripe)
Multi-tenant SaaS API
File Storage Service (S3-like)
Notification Service (Email/SMS/Push)
API Gateway for Microservices
Social Media Backend (Feed, Comments, Likes)
Stuck on this topic? Ask an Insider
Get 1:1 guidance from people who've walked this exact path — on the InsideEdge app.
