DevServe

Jan 2026 — Present · View on GitHub →

A CLI/daemon tool that manages local dev server processes and exposes them via Tailscale Serve for remote access. Also includes a TUI that easily manages processes.

Tech Stack

• Go
• Cobra
• Tailscale
• Bubbletea

What the heck is this

For work, I tend to have a fair amount of dev servers running while working on front end projects. This lead to a lot of headaches when I wanted to view some changes on my phone while eating lunch, and even while stopping and starting these processes, I frequently ran into the problem of the ports being used by ghost tasks. I had enough and built devserve, which helps me manage my dev servers. I also use tailscale to run these devservers on my home PC instead of my work laptop so I can have them running constantly. Tailscale allows me to view these sites on devices connected to my tailnet.

My workflow

Since my home PC is much faster than my work laptop, I tend to ssh into the PC and develop my projects there. I needed a way to view these sites, and I chose tailscale. This also helped facilitate the ssh process as a whole, and it is more secure than port forwarding specific ports and doing all that nonsense. For developing projects this is a much more seamless process.

How to use it

First you need a project. Most of mine are NextJS projects, which makes it quite simple. Then you run the dev server using devserve

devserve serve myapp 3000 "yarn dev --port 3000"

Now the devserve daemon is tracking this process. I can add it to a global config file to easily spin it up later just based on the name of the process

devserve start myapp

Other motivations

I also realized that sometimes AI agents like to spin up these dev servers, and when they do, they tend to get stuck. I implemented a feature where the logs of the processes tracked by devserve are output to a file. This will allow your agents to read the logs without having ownership of the process, and makes the debugging process a lot easier. Just put in a skill or your agents.md file a line that explains to read the logs in the .devserve directory and your agent will be able to handle debugging already running dev servers much easier.