EVE BOT

A team of always-on AI teammates. Each one has its own computer, works inside the apps you already use, and keeps going after you close the laptop, coming back only when something needs your approval.

Where do you want it to run?

On Vercel

One click. Vercel runs the computer, the models, the storage, and the schedules, and bills your account.

  • Copies this repository to your GitHub, creates the project, and deploys it.
  • Asks for one thing: a console password. openssl rand -base64 32 makes a good one.
  • Then, in the project's Storage tab, create a Blob store and connect it. It holds the roster, jobs, and memory. Redeploy once.
  • The team's computer is a Vercel Sandbox; models go through AI Gateway.
Deploy to Vercel

Open /bot on your deployment and sign in with the password; the setup page there checks each step off. Hobby plans work; Pro runs the watchdog every minute.

On your own machine

Nothing leaves your network unless you point it there. The computer runs in Docker, models come from Ollama, LM Studio, OpenRouter, or any OpenAI-compatible API, and storage is on disk.

You need Node 24, Docker, and a model endpoint.

git clone https://github.com/Shaivpidadi/eve-bot.git && cd eve-bot
npm install
npm run setup        # pick "Standalone", answer a few questions
npm run build && npm start
# open http://localhost:3000/bot

Or as containers, with a SearXNG for web search and an optional Ollama alongside:

npm run setup
docker compose up -d --build
Read the standalone guide

Either way, this is what you get

HQ and a rosterMessage HQ like a colleague. It writes a brief, picks the right Bot, and hands off the job.
A computer the team keepsOne real Chrome, a terminal, and files, shared by every Bot and backed up while they work.
Watch and take overThe screen sits in every thread. Sign-ins, codes, and CAPTCHAs are yours; passwords never touch chat.
Durable jobsWork survives restarts and redeploys, waits for its start time without compute, and asks before anything irreversible.