your localhost,
on the public internet.

expose localhost to the internet instantly. inspect every request, replay it with one key. no account, no rate limits, self-hosted.

read docs
or
zero configno signup100% private
custom 9-byte binary multiplexer/
< 1ms frame latency/
zero cloud lock-in/
bubble tea terminal tui/
instant 1-key webhook replay/
custom subdomains/
shared token authentication/
100% private self-hosted/
dual stream telemetry/
0 allocation fast-path/
custom 9-byte binary multiplexer/
< 1ms frame latency/
zero cloud lock-in/
bubble tea terminal tui/
instant 1-key webhook replay/
custom subdomains/
shared token authentication/
100% private self-hosted/
dual stream telemetry/
0 allocation fast-path/
live terminal inspection

real-timetelemetrywithoutleavingyourshell

every webhook and http request streaming through your multiplexed tunnel, right in bubble tea tui.

tunneru 3000
tunnel:https://myapp.tunneru.knightkun.codes
forwarding:http://localhost:3000
inspect:http://localhost:4040
connected
time
method
path
status
lat
18:24:02
POST
/api/webhooks/stripe
200 ok
14ms
18:24:08
GET
/api/v1/user/profile
200 ok
8ms
18:24:15
POST
/webhooks/github/push
200 ok
19ms
18:24:22
POST
/api/auth/token
401 unauthorized
12ms
4 requests3 ok1 err13ms avg
[c] clear[↑/↓] scroll[q] quit
engineering highlights

builtfordeveloperswhovaluespeed&privacy

every layer engineered from the ground up without third-party cloud lock-in or heavy runtime overhead.

custom 9-byte binary multiplexer

multiplex hundreds of concurrent http streams over a single persistent tcp socket. framing is achieved with an ultra-compact 9-byte binary header, eliminating websocket and http/2 protocol overhead.

< 1ms
multiplex latency
9 bytes
frame overhead
0 allocs
fast-path parsing

1-click webhook replay

fixed a bug in your webhook handler? replay the captured payload directly from the inspector without triggering external stripe or github events again.

POST /api/webhookreplay

bubble tea tui & web ui

rich terminal interface with live requests and latency metrics, paired with a web inspector for formatted json diffs and raw headers.

terminal: live tuiweb: localhost:4040

subdomains & token auth

claim predictable subdomains with --subdomain. lock down access with shared token authentication.

$ tunneru 3000 -s myapp -t token

zero signup & 100% private

no email required, no cloud accounts, and no request logs stored on external servers. pure open source developer infrastructure.

open sourcemit licensed
architecture & data flow

howtunneruroutestrafficinmicroseconds

a custom binary multiplexing protocol built from scratch in go for zero latency and minimal cpu overhead.

01

public ingress & routing

https://myapp.tunneru.knightkun.codes

external traffic hits the public edge. the reverse proxy inspects the host header, looks up the active client session in the concurrent registry, and locks a stream.

protocolhttp/1.1 & https
subdomainmyapp.tunneru.knightkun.codes
lookup< 1ms registry match
02

9-byte binary multiplexer

persistent tcp stream

raw http requests are framed into 9-byte header binary frames (type, stream id, length, payload) and streamed over a single persistent tcp socket with zero backpressure delays.

frame header1B type + 4B id + 4B len
sessionbidirectional full-duplex
framingzero serialization overhead
03

local dispatch & telemetry

localhost:3000 + :4040

the local tunneru daemon unwraps the frames, forwards the http request to your local dev server, and simultaneously broadcasts the request record to the tui and web dashboard.

targethttp://localhost:3000
terminalbubble tea live telemetry
inspectorwebsocket event hub (:4040)
wire format: 9-byte multiplexer frame header
offset 0x00 (1 byte)
frame type
0x01 data / 0x02 ping / 0x04 close
offset 0x01 (4 bytes)
stream id
uint32 big-endian stream id
offset 0x05 (4 bytes)
payload length
uint32 big-endian payload size
offset 0x09 (n bytes)
payload data
raw chunked http frame bytes
developer guide & reference

everythingyouneedtogetstarted

install the client, manage tokens, inspect live traffic, or run your own server.

1. install the client

download and install the prebuilt static binary for macos, linux, or windows.

bash
curl -fsSL https://raw.githubusercontent.com/pranav718/tunneru/main/install.sh | sh

2. expose a local port

run your local application (e.g. next.js, fastApi, flask, rails) on port 3000, then start the tunnel.

bash
tunneru 3000

tunneru assigns a random public subdomain (e.g. https://frost-729.tunneru.knightkun.codes) and starts the live tui.

3. request a custom subdomain

specify a desired subdomain if available on the server.

bash
tunneru 3000 --subdomain myapp
read full protocol spec & architecture in github readme