MCP servers, pluggable per bot
Point a bot at any Model Context Protocol server over HTTP or SSE. Tools are discovered automatically and you pick which bots may call which tools.
Bring your own API key, paste a bot token, start talking. No server to run, nothing to install.
ops assistant
bot ยท node build-01 online
Works with the keys you already have
Anthropic, OpenAI, Gemini, OpenRouter or any OpenAI-compatible endpoint. Encrypted at rest, tested on save.
Create a bot with BotFather, paste the token. We verify it and wire the webhook. Add as many bots as you like.
Your first message gets a pairing code. Approve it once in the dashboard and the bot is yours. Nobody else gets in.
Everything runs on our side until you decide to plug in your own tools, servers or machines.
Point a bot at any Model Context Protocol server over HTTP or SSE. Tools are discovered automatically and you pick which bots may call which tools.
Persona, notes and daily memory live as plain files you can read and edit.
One command connects Windows, macOS, Linux or a Docker host. Files and shell, with approvals.
A terse ops bot on one model, a chatty research bot on another. Each with its own prompt and access list.
We never resell tokens. Usage shows up on your provider invoice, and in the dashboard.
A node is a small agent you run anywhere. It opens an outbound connection, declares what it can do, and waits. Risky commands ask for approval right in the chat.
# on your machine
npx easylopster-node connect 7KQ2-M9XA
# or as a container with its own filesystem
docker run -d --name el-node \
-e EASYLOPSTER_TOKEN=... \
-v el-workspace:/workspace \
ghcr.io/noneym/easylopster-nodeThree bots, two nodes, five MCP servers per account. Enough to run a small team's worth of assistants.