Skip to content

Latest commit

 

History

History
41 lines (26 loc) · 1.2 KB

README.md

File metadata and controls

41 lines (26 loc) · 1.2 KB

ipfs-connector

Using the IPFS to transfer files between computers can be slow because it can take time to locate a path between the two computers on the IPFS network. This makes it faster by lowering the barrier towards making sure your computers are in each other's swarm.

First run the IPFS daemon on your computers

$ ipfs daemon

Then get ipfs-connect using go:

$ go get github.com/schollz/ipfs-connect

Then connect the two computers. On the first computer type:

$ ipfs-connect mycomps
your id: jkljl88-ji98-449-a0e1-c87a04802922
add another computer to your swarm by running

ipfs-connect mycomps

And then on the other computer type

$ ipfs-connect mycomps

And then, voila! Your computers will be swarmed together as long as the IP addresses don't change. Now you can share files via IPFS between two computers without waiting.

How does it work?

This uses a simple rendezvous server: schollz/duct which is a ephemeral MPMC pubsub. Both computers connect to the duct tape server and listen for a payload about their IPFS addresses. Once they receive the addresses they connect to them via the ipfs swarm connect command.

License

MIT