Skip to content
Tarisu Limited

Bot hosting

Your bot keeps running when your laptop does not. Discord, Telegram, Matrix and Slack, four plans in Istanbul (IST), 512 MB to 4 GB of RAM and 50% to 200% CPU, built in seconds when your payment clears and run from a web panel: console, file manager, restart.

Choose the right server

By quote

A custom quote for Bot Hosting

We configure and price this service to fit your needs. Tell us the number of accounts, your domain and what you expect, and we will prepare a quote for you.

What happens after you order

  1. Step 1

    Pick a plan and a runtime

    Choose the size, then which network you are writing for: Discord, Telegram, Matrix or Slack, and in which language. An invoice is created for the order.

  2. Step 2

    Pay the invoice

    Pay with one of the methods offered at checkout. Your order is accepted once the payment is recorded.

  3. Step 3

    The container is built in seconds

    It arrives with the runtime you chose already installed, and the panel login is sent to the e-mail address on your account.

  4. Step 4

    Upload your bot and start it

    Put the files in through the panel file manager or over SFTP, enter your token or access key, set the start command and press start. The console shows the output live.

A process that stays running

A bot on your laptop stops when the laptop closes, updates or drops its connection. Moving the process onto a machine in Istanbul that is always on is the whole purchase, and it is the same purchase whether the bot answers in a Discord guild, a Telegram group, a Matrix room or a Slack channel.

The container is built when your payment clears, in seconds, and everything after that happens in the web panel: upload the files, set the start command, watch the console, restart.

No shell, no root and no operating system of your own; that is the trade, and a VPS is where you go if you need the other side of it.

  • Always on, independent of your machine
  • Discord, Telegram, Matrix and Slack
  • No shell and no root: that is a VPS

Choosing the size of a plan

One question decides it: is the bot holding things in memory or chewing on audio.

Member caches for busy guilds and a database write on every message are memory problems, and they grow with how active the rooms are rather than with how many there are. Voice and music are processor problems, because the stream is decoded and re-encoded for as long as someone is listening, and they only arise on the Discord side.

The plans run 512 MB to 4 GB of RAM and 50% to 200% CPU, where 100% is one core, so pick the side you are on and read down that column.

  • Memory for caches, processor for audio
  • 512 MB to 4 GB RAM, 50% to 200% CPU
  • 100% CPU is one core

What comes with Bot Hosting

  • Discord, Telegram, Matrix and Slack
  • Web panel access
  • Reserved CPU and memory
  • Console, restart and reinstall
  • Weekly backup
  • Istanbul (IST)

Four networks, sixteen ways to write for them

  • Discord · Node.js (discord.js)
  • Discord · Python (discord.py)
  • Discord · Bun (TypeScript / JavaScript)
  • Discord · Deno (TypeScript)
  • Discord · Elixir (Nostrum)
  • Discord · Go (discordgo)
  • Discord · Java (JDA / Javacord / Kord)
  • Discord · Lua (Discordia / Luvit)
  • Discord · Ruby (discordrb)
  • Discord · Rust (serenity / twilight)
  • Telegram · Python (python-telegram-bot)
  • Telegram · Node.js (telegraf)
  • Matrix · Python (matrix-nio)
  • Matrix · Node.js (matrix-bot-sdk)
  • Slack · Python (Bolt)
  • Slack · Node.js (Bolt)

Who is Bot Hosting for?

  • Moderation and community bots

    Logging actions, handing out levels and roles, writing to a database on almost every message. The commonest job on the Discord and Telegram side, and a memory one.

  • Notification and alert relays

    Polling a feed or an API and posting into a Telegram or Slack channel costs little memory, just a process that stays connected.

  • Bots for your own Matrix server

    Bridges, archivers and welcome bots connect to your homeserver. The default is matrix.org and you can point it at your own address.

  • Voice and music bots

    Audio is bound by processor rather than memory, so size these on the CPU percentage. A Discord-side job.

What Bot Hosting is like to use

Every runtime on every plan

Discord in Node.js, Python, Bun, Deno, Elixir, Go, Java, Lua, Ruby and Rust; Telegram, Matrix and Slack in Python and Node.js. All sixteen are rebuild targets on all four sizes, so the language is a decision you can change later from the panel. Changing it rebuilds the container from the new runtime and deletes what is on it, which is what the panel warns before it runs, so take a copy of your files first.

A MySQL database, not a file

One database per plan, created with the container. Levels, warnings, scores and seen-at timestamps belong in it rather than in a JSON file that a crash can truncate halfway through a write.

A port of your own

One allocation on every plan, which is what a webhook listener, an OAuth callback or a health endpoint needs. A bot that only opens outbound connections never touches it, and loses nothing by having it.

Restarted, not watched

When the process exits the container starts it again, and the console keeps the output from before it went. A crash loop shows up in the panel as a crash loop, rather than as a quiet absence somebody reports to you hours later.

CPU counted in whole cores

50% is half a Xeon E5v4 core, 100% is one and 200% is two. Text commands rarely reach the first of those; voice and audio encoding is the work that does, and it is the reason to read this line before the memory one.

In Istanbul, from $0.99 a month

One location, so there is no choice to get wrong. For a community that is mostly in Turkey, the round trip to the gateway and back is a local one, and the smallest plan costs less than the domain the bot is named after.

Frequently asked questions

Could not find your answer? Write to us and we will route your question to the right team.

Still unanswered
  • Which bot platforms can I host?

    Four: Discord, Telegram, Matrix and Slack. They run in the same container, in the same panel, at the same price, and the only difference is the runtime you pick at the order screen. Telegram, Matrix and Slack come in Python and Node.js; Discord comes in ten languages.

  • Can I host a Telegram bot?

    Yes. python-telegram-bot, aiogram and pyTelegramBotAPI on the Python side, telegraf, grammY and node-telegram-bot-api on the Node.js side. You put your bot token in from the panel and upload your code through the file manager or over SFTP.

  • Can I point a Matrix bot at my own homeserver?

    Yes. The homeserver address is a setting: it defaults to matrix.org and you can write your own in. Authenticate with an access token or a password. matrix-nio is installed on the Python side and matrix-bot-sdk on the Node.js side.

  • How much RAM does a bot need?

    A small command bot runs in tens of megabytes. One holding a member cache for busy Discord guilds needs hundreds, and that grows with how active the guilds are rather than with how many there are. Music and voice are a processor question rather than a memory one. The plans run 512 MB to 4 GB, so start at the size that matches the side you are on.

  • Can I run bots for more than one network on a single plan?

    A container is built with one runtime, but what you run inside it is yours: one Python container can hold a Telegram bot and a Matrix bot. If you want two different languages, take a second plan.

  • Does my bot stay online all the time?

    That is the whole purchase. The container runs on a machine in Istanbul that is always on, independently of your own computer, and restarts on its own if the process exits.

  • How do I upload my bot files?

    Through the file manager in the panel, or over SFTP with the credentials the panel gives you. Then set the start command and press start; the console shows the output live.

  • Is there a backup on bot hosting?

    Yes, a weekly backup on every plan, and you restore it yourself from the panel. The number of backup slots a plan holds is in its configuration.

  • How many bots can I run on one plan?

    As many as the memory allows. Nothing in the container enforces a count and we do not sell one. A small bot needs tens of megabytes; one holding a member cache for a busy guild needs hundreds, so the ceiling is the RAM and CPU on your plan rather than a rule of ours.

  • What happens if my bot crashes?

    The container restarts the process, and the console keeps the output that led up to it so you can see what happened. A crash loop is visible in the panel rather than something you find out about from your users.

  • What does this give me that a VPS does not?

    A panel instead of a machine, so there is no operating system to install, patch or secure. The limit is the other side of the same thing: no root access and no shell. A bot that needs its own system service, or a database engine beside it, belongs on a VPS.

  • Can I host a bot that breaks the platform rules?

    No. Bots that violate the terms of the network they run on, or our service agreement, including spam, self-bots and unauthorised data collection, are suspended when we find them. If what your bot does sits near that line, ask us before you order rather than after.

Size it by what the bot remembers

A bot that answers a command and forgets it runs in tens of megabytes, so for that one the smallest plan is not a starter tier, it is the right tier. What moves a bot up this list is holding things: a member cache for busy guilds, a queue, an audio stream. Moving up later keeps your files and your runtime, so starting small costs nothing but the move.

We use strictly necessary cookies to keep you signed in and remember your preferences. We would also like to measure how the site is used, but only with your agreement. Read our privacy policy