A local, Netflix-style streaming library built with Streamlit.
Drop video files in a folder, browse them as a poster grid, and play them in the browser. You can use this any way you like, but launching this app and running it as a local hosted web app allows you to stream video to any device in combination with a VPN.
You can use it solely on your PC alone, but with a VPN like Tailscale to create allowed devices to connect to, you can connect remotely on any device with a browser.
Launch this app on a home computer, use a VPN like Tailscale to connect to your home network from a remote device, you can then use the host's VPN address to open the app on your remote device's browser, allowing you to create your own private streaming platform with your home computer as your own private server.
The URL to connect on your phone or other device should look like:
100.xxx.xxx.xx:8501
This is intended for private use only. Copyright law still applies.
SCREEN 01
Now Showing
The main browse view. This is a mockup reflex recreation of the app. A poster grid of everything in the library, with a sidebar for search and genre filters, plus quick access to add or manage movies.
Browse the collection
All Movies
SCREEN 02
Add a Movie
Adding to the library. Movies can be pulled in by scanning the local movies folder, importing a whole directory, or entering a single film by hand.
Scan Movies Folder
Check the local 'movies' folder for new or removed videos, and update the library.
Add a Directory
Add a Single Movie
Movie Details
Preview
Title: —
Path: —
Poster: —
Genre: Action
SCREEN 03
Manage Library
Library maintenance. Every entry listed with genre and file path, a search box, and per-row checkboxes for editing or bulk-removing entries.
4 of 10 movies shown.
Dune: Part Two
Science Fiction · D:\\Videos\\Dune: Part Two.mkv
Blade Runner 2049
Science Fiction · D:\\Videos\\Blade Runner 2049.mkv
The Batman
Action · D:\\Videos\\The Batman.mkv
Interstellar
Science Fiction · D:\\Videos\\Interstellar.mkv
SCREEN 04
Video Player
The playback view. MovieFlex streams video straight from disk with HTTP range requests, so seeking works without loading the whole file into memory.
Genre: Science Fiction
Playback preview · the real app streams the full file from disk.
SCREEN 05
About MovieFlex
What the project is, what it's built with, and where the code lives.
Let's get technical. How does this work?
How it works
MovieFlex is a local movie library and streaming app built with Streamlit. It reads a JSON catalog, displays the collection as a poster grid, and streams video files directly from disk with full seek support. PyInstaller packages it as a standalone Windows release.
Navigation runs on Streamlit's session state. st.session_state is per-browser-session memory that survives reruns, so it remembers which view is open and which movie is playing. In MovieFlex, a single active_video value decides which view function runs. Clicking a button updates that value and reruns the script, switching to the Add, Manage, Edit, or player view. Each view has its own logic, but they all share the same session.
Project map
MovieFlexApp.pyStreamlit UI, catalog, and range-streaming playerlauncher.pyStarts Streamlit on port 8501DataBase.jsonMovie catalog and metadatatheme.cssCustom app stylingMovieFlex.specPyInstaller packaging configThe data layer
The catalog is DataBase.json. Each entry stores a file path, title, poster, and genre, read and written as a Python dictionary.
Posters can also be discovered from sidecar images next to the video. Manual entries are straightforward; the directory scanner uses Python's os module to compare normalized file paths against the database and add missing files. It also removes scan-sourced entries whose files have disappeared while leaving manual entries alone.
What I learned
It's also worth noting that Streamlit's limitations caused several roadblocks. The biggest was video playback. The built-in player struggled with large local files, so huge movies were out of the question. Streamlit runs on port 8501, so I added a small HTTP server class on port 8502 that streams video directly from disk instead.
I learned a lot from this one. It was my first project using this kind of data-driven functionality. I'm not that big into movies, but I guess I have my own private Netflix now.
Built with
Playback pipeline
| Catalog | DataBase.json |
| Views | Streamlit session state |
| Player | HTTP range server :8502 |
| Build | PyInstaller executable |
(GitHub link placeholder — will point to the repo once git-hub solves my ticket.)
GitHub repositoryPart of the coding collection.