https://github.com/ambianic/peerfetch

Peer-to-peer HTTP over WebRTC.
https://github.com/ambianic/peerfetch

end-to-end-encryption iot webrtc

Last synced: about 2 months ago
JSON representation

Peer-to-peer HTTP over WebRTC.

• Host: GitHub
• URL: https://github.com/ambianic/peerfetch
• Owner: ambianic
• License: apache-2.0
• Created: 2021-08-15T03:34:42.000Z (about 3 years ago)
• Default Branch: main
• Last Pushed: 2023-03-14T18:12:37.000Z (over 1 year ago)
• Last Synced: 2024-05-01T10:02:28.592Z (5 months ago)
• Topics: end-to-end-encryption, iot, webrtc
• Language: TypeScript
• Homepage:
• Size: 1.4 MB
• Stars: 61
• Watchers: 1
• Forks: 7
• Open Issues: 10
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-webrtc - peerfetch - Peer-to-peer HTTP over WebRTC. (JavaScript)

README

        

# peerfetch

Peer-to-peer HTTP over WebRTC. Implements an http client wrapper (similar to [HTML fetch](https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API/Using_Fetch)) and a corresponding server side HTTP proxy over WebRTC DataChannel. Allows direct secure access from a web browser to edge devices (IoT or servers) hidden behind a firewall. 

Highlights:

- Excutes in safe application space (no `sudo`)

- End to end encryption

- No custom VPN setup needed

- No cloud middleman for data tunneling

- No dynamic DNS service required

- No custom firewall rules required

- Programming language agnostic. (Currently available in JavaScript and Python)

# Hello World example

See this [README](examples/helloworld/README.md) for a step by step Hello World example.

# How it works

For an in-depth technical discussion and project background, see [this blog post](https://webrtchacks.com/private-home-surveillance-with-the-webrtc-datachannel/).

# Hands-on Workshop

[Here is the recording](https://www.youtube.com/watch?v=LFKYtL1_RjQ) of a `peerfetch` workshop hosted by [Python Austin](https://github.com/Jacob-Barhak/EveningOfPythonCoding).

# Use cases:

- Direct user access from a web app to private home security camera without sharing footage with a cloud provider.

- IoT device mesh with direct p2p communication. 

- Personal web apps can share data directly (files, notes, photos) only with the end user without exposing a public IP address.

- Federated learning - ML models can train on local user data and share learned states directly with each other without a centralized model aggregation server.

# Used by

- [Ambianic UI PWA](https://github.com/ambianic/ambianic-ui)

- [Ambianic Edge](https://github.com/ambianic/ambianic-edge)