5 releases

0.1.19 Jul 19, 2021
0.1.17 May 16, 2021
0.1.16 May 1, 2021
0.1.14 Apr 18, 2021
0.1.10 Jul 13, 2020

#1036 in HTTP server

MIT license

8KB
179 lines

BuildRelease crate GitHub Docker Registry crate

tunnelto

tunnelto lets you expose your locally running web server via a public URL. Written in Rust. Built completely with async-io on top of tokio.

  1. Install
  2. Usage Instructions
  3. Host it yourself

Install

Brew (macOS)

brew install agrinman/tap/tunnelto

Cargo

cargo install tunnelto

Everywhere

Or Download a release for your target OS here: tunnelto/releases

Usage

Quick Start

tunnelto --port 8000

The above command opens a tunnel and forwards traffic to localhost:8000.

More Options:

tunnelto 0.1.14

USAGE:
    tunnelto [FLAGS] [OPTIONS] [SUBCOMMAND]

FLAGS:
    -h, --help       Prints help information
    -V, --version    Prints version information
    -v, --verbose    A level of verbosity, and can be used multiple times

OPTIONS:
        --dashboard-address <dashboard-address>    Sets the address of the local introspection dashboard
    -k, --key <key>                                Sets an API authentication key to use for this tunnel
        --host <local-host>
            Sets the HOST (i.e. localhost) to forward incoming tunnel traffic to [default: localhost]

    -p, --port <port>
            Sets the port to forward incoming tunnel traffic to on the target host

        --scheme <scheme>
            Sets the SCHEME (i.e. http or https) to forward incoming tunnel traffic to [default: http]

    -s, --subdomain <sub-domain>                   Specify a sub-domain for this tunnel

SUBCOMMANDS:
    help        Prints this message or the help of the given subcommand(s)
    set-auth    Store the API Authentication key

Host it yourself

  1. Compile the server for the musl target. See the musl_build.sh for a way to do this trivially with Docker!
  2. See Dockerfile for a simple alpine based image that runs that server binary.
  3. Deploy the image where ever you want.

Testing Locally

# Run the Server: xpects TCP traffic on 8080 and control websockets on 5000
ALLOWED_HOSTS="localhost" cargo run --bin tunnelto_server

# Run a local tunnelto client talking to your local tunnelto_server
CTRL_HOST="localhost" CTRL_PORT=5000 CTRL_TLS_OFF=1 cargo run --bin tunnelto -- -p 8000

# Test it out!
# Remember 8080 is our local tunnelto TCP server
curl -H '<subdomain>.localhost' "http://localhost:8080/some_path?with=somequery"

See tunnelto_server/src/config.rs for the environment variables for configuration.

Caveats for hosting it yourself

The implementation does not support multiple running servers (i.e. centralized coordination). Therefore, if you deploy multiple instances of the server, it will only work if the client connects to the same instance as the remote TCP stream.

The version hosted by us is a proper distributed system running on the the fabulous fly.io service. In short, fly.io makes this super easy with their Private Networking feature. See tunnelto_server/src/network/mod.rs for the implementation details of our gossip mechanism.

Dependencies

~2.5–3.5MB
~70K SLoC