Rathole, like frp and ngrok, can help to expose the service on the device behind the NAT to the Internet, via a server with a public IP.


  • High Performance Much higher throughput can be achieved than frp, and more stable when handling a large volume of connections. See Benchmark
  • Low Resource Consumption Consumes much fewer memory than similar tools. See Benchmark. The binary can be as small as ~500KiB to fit the constraints of devices, like embedded devices as routers.
  • Security Tokens of services are mandatory and service-wise. The server and clients are responsible for their own configs. With the optional Noise Protocol, encryption can be configured at ease. No need to create a self-signed certificate! TLS is also supported.
  • Hot Reload Services can be added or removed dynamically by hot-reloading the configuration file. HTTP API is WIP.


A full-powered rathole can be obtained from the release page. Or build from source for other platforms and minimizing the binary. A Docker image is also available.

The usage of rathole is very similar to frp. If you have experience with the latter, then the configuration is very easy for you. The only difference is that configuration of a service is split into the client side and the server side, and a token is mandatory.

To use rathole, you need a server with a public IP, and a device behind the NAT, where some services that need to be exposed to the Internet.

Assuming you have a NAS at home behind the NAT, and want to expose its ssh service to the Internet:

  1. On the server which has a public IP

Create server.toml with the following content and accommodate it to your needs.

bind_addr = “” # 2333 specifies the port that rathole listens for clients
token = “use_a_secret_that_only_you_know” # Token that is used to authenticate the client for the service. Change to a arbitrary value.
bind_addr = “” # 5202 specifies the port that exposes my_nas_ssh to the Internet

Then run:

./rathole server.toml

  1. On the host which is behind the NAT (your NAS)

Create client.toml with the following content and accommodate it to your needs.


remote_addr = “myserver.com:2333” # The address of the server. The port must be the same with the port in server.bind_addr

token = “use_a_secret_that_only_you_know” # Must be the same with the server to pass the validation
local_addr = “” # The address of the service that needs to be forwarded

  1. Now the client will try to connect to the server myserver.com on port 2333, and any traffic to myserver.com:5202 will be forwarded to the client’s port 22.

So you can ssh myserver.com:5202 to ssh to your NAS.

To run rathole run as a background service on Linux, checkout the systemd examples.


rathole can automatically determine to run in the server mode or the client mode, according to the content of the configuration file, if only one of [server] and [client] block is present, like the example in Quickstart.

But the [client] and [server] block can also be put in one file. Then on the server side, run rathole --server config.toml and on the client side, run rathole --client config.toml to explicitly tell rathole the running mode.

Before heading to the full configuration specification, it’s recommend to skim the configuration examples to get a feeling of the configuration format.

See Transport for more details about encryption and the transport block.

Here is the full configuration specification

remote_addr = “example.com:2333” # Necessary. The address of the server
default_token = “default_token_if_not_specify” # Optional. The default token of services, if they don’t define their own ones
heartbeat_timeout = 40 # Optional. Set to 0 to disable the application-layer heartbeat test. The value must be greater than server.heartbeat_interval. Default: 40 seconds
[client.transport] # The whole block is optional. Specify which transport to use
type = “tcp” # Optional. Possible values: [“tcp”, “tls”, “noise”]. Default: “tcp”
[client.transport.tcp] # Optional. Also affects noise and tls
proxy = “socks5://user:passwd@” # Optional. The proxy used to connect to the server. http and socks5 is supported.
nodelay = false # Optional. Determine whether to enable TCP_NODELAY, if applicable, to improve the latency but decrease the bandwidth. Default: false
keepalive_secs = 20 # Optional. Specify tcp_keepalive_time in tcp(7), if applicable. Default: 20 seconds
keepalive_interval = 8 # Optional. Specify tcp_keepalive_intvl in tcp(7), if applicable. Default: 8 seconds
[client.transport.tls] # Necessary if type is “tls”
trusted_root = “ca.pem” # Necessary. The certificate of CA that signed the server’s certificate
hostname = “example.com” # Optional. The hostname that the client uses to validate the certificate. If not set, fallback to client.remote_addr
[client.transport.noise] # Noise protocol. See docs/transport.md for further explanation
pattern = “Noise_NK_25519_ChaChaPoly_BLAKE2s” # Optional. Default value as shown
local_private_key = “key_encoded_in_base64” # Optional
remote_public_key = “key_encoded_in_base64” # Optional
[client.services.service1] # A service that needs forwarding. The name service1 can change arbitrarily, as long as identical to the name in the server’s configuration
type = “tcp” # Optional. The protocol that needs forwarding. Possible values: [“tcp”, “udp”]. Default: “tcp”
token = “whatever” # Necessary if client.default_token not set
local_addr = “” # Necessary. The address of the service that needs to be forwarded
nodelay = false # Optional. Determine whether to enable TCP_NODELAY for data transmission, if applicable, to improve the latency but decrease the bandwidth. Default: false
[client.services.service2] # Multiple services can be defined
local_addr = “”
bind_addr = “” # Necessary. The address that the server listens for clients. Generally only the port needs to be change.
default_token = “default_token_if_not_specify” # Optional
heartbeat_interval = 30 # Optional. The interval between two application-layer heartbeat. Set to 0 to disable sending heartbeat. Default: 30 seconds
[server.transport] # Same as [client.transport]
type = “tcp”
[server.transport.tcp] # Same as the client
nodelay = false
keepalive_secs = 20
keepalive_interval = 8
[server.transport.tls] # Necessary if type is “tls”
pkcs12 = “identify.pfx” # Necessary. pkcs12 file of server’s certificate and private key
pkcs12_password = “password” # Necessary. Password of the pkcs12 file
[server.transport.noise] # Same as [client.transport.noise]
pattern = “Noise_NK_25519_ChaChaPoly_BLAKE2s”
local_private_key = “key_encoded_in_base64”
remote_public_key = “key_encoded_in_base64”
[server.services.service1] # The service name must be identical to the client side
type = “tcp” # Optional. Same as the client [client.services.X.type] token = "whatever" # Necessary ifserver.default_token` not set
bind_addr = “” # Necessary. The address of the service is exposed at. Generally only the port needs to be change.
nodelay = false # Optional. Same as the client
bind_addr = “”


rathole, like many other Rust programs, use environment variables to control the logging level. infowarn, errordebugtrace are available.

RUST_LOG=error ./rathole config.toml


rathole has similar latency to frp, but can handle a more connections, provide larger bandwidth, with less memory usage.

For more details, see the separate page Benchmark.

However, don’t take it from here that rathole can magically make your forwarded service faster several times than before. The benchmark is done on local loopback, indicating the performance when the task is cpu-bounded. One can gain quite a improvement if the network is not the bottleneck. Unfortunately, that’s not true for many users. In that case, the main benefit is lower resource consumption, while the bandwidth and the latency may not improved significantly.

Development Status

rathole is under active development. A load of features is on the way:

  •  TLS support
  •  UDP support
  •  Hot reloading
  •  HTTP APIs for configuration

Leave a comment

Your email address will not be published. Required fields are marked *