Seven options side by side
| Price a month | When it stops | Long polling | Webhook | |
|---|---|---|---|---|
| Your laptop | free | When the lid closes or the Wi-Fi drops | works while it is on | needs a public https address |
| Render, free web service | free | Spins down after 15 minutes without inbound traffic | no inbound traffic, so it spins down | works; the first message waits about one minute |
| Heroku Eco | $5, 1,000 dyno hours a month shared by all Eco dynos in an account | An Eco web dyno that gets no web traffic for 30 minutes sleeps; web traffic wakes it after a short delay. An Eco worker dyno does not sleep, because it does not answer web requests | works on a worker dyno | works; web traffic wakes a sleeping dyno |
| Serverless function (AWS Lambda) | free tier: 1 million requests and 400,000 GB-seconds a month | Each call ends after 15 minutes at most; built for short tasks that keep no state between calls | does not fit | one call per message |
| Container platform (Fly, Railway) | $6.70–30 | Runs and bills while it runs | works | works |
| VPS (Hetzner, DigitalOcean) | $4.5–6.5 budget, $5–7 cloud | Runs and bills all month | works | works |
| Uno computer | from $5 | Sleeps after 10 minutes with nothing to do | does not wake the computer | each message wakes the computer |
What it costs to keep one small bot answering
Plan price for about 1 vCPU and 1 GB, in $ per month. Ranges show the cheapest and the dearest provider we checked.
Show as a table
| Option | $ per month | Shape |
|---|---|---|
| Uno Small | 5 | 1 vCPU · 1 GB · sleeps when idle, wakes on a webhook message |
| Render free | 0 | Free, spins down after 15 minutes without inbound traffic |
| Heroku Eco | 5 | 512 MB · a worker dyno does not sleep |
| Heroku Basic | 7 | 512 MB · no sleeping |
| Budget VPS | 4.5–6.5 | Hetzner CX, OVH, Contabo · runs and bills all month |
| Cloud VPS | 5–7 | DigitalOcean, Vultr, Linode, Lightsail |
| Container platforms | 6.70–30 | Fly shared 1 GB up to Railway at full load |
Polling or webhook decides which hosts fit
A polling bot asks Telegram for news in a loop. All of its traffic goes out, so a sleeping host never hears anything. A webhook bot waits for Telegram to call its https address, and that call is the request that wakes a sleeping host. Telegram calls webhooks on ports 443, 80, 88 and 8443, over TLS 1.2 or newer; older versions are not supported.
That is why the sleeping hosts split. Render's free service and Heroku's web dyno wake on web traffic, so a webhook bot fits them. A polling bot fits a worker dyno, a container platform or a VPS.
Which one for which bot
| Your bot | Take | Why |
|---|---|---|
| Polling, answers all day | VPS or a Heroku Eco worker | Both keep running. A Heroku worker uses about 744 of the 1,000 Eco hours in a 31-day month |
| Webhook, quiet most of the day | Uno, or a free tier you can live with | The rule on Uno: a bot on a webhook wakes the computer with each message; a bot on long polling cannot, so switch it to a webhook |
| A few calls a day, no state | A serverless function | The free tier covers 1 million requests and 400,000 GB-seconds a month; nothing is kept between calls |
| A bot that keeps files or a database | Uno or a VPS | Disk and memory stay with the computer; Uno adds 25 GB of cloud storage on Small |
Where Uno is worse
- Long polling does not fit a sleeping host: a bot on long polling only makes outgoing requests, and an outgoing request does not wake a sleeping computer. Use a webhook, or take a VPS.
- The first message after a sleep waits for the wake-up. Our measurement of 3 October 2026 gave a median 118 ms through the console.
- A Heroku Eco worker costs $5 and never sleeps. Uno Small costs $5 and sleeps. If the bot polls, the worker is the simpler fit.
- Budget VPS hosts sell memory cheaper: $0.65–1.6 per GB a month against $4.2–5 on Uno.
- An agent builds the bot for you only from Plus: on Small the computer is a bare server for your own agent; Uno Work in the cloud starts with Plus.
- Paid Uno computers run in the Netherlands today. Telegram reaches them from anywhere.
Uno against a classic VPS for a bot
The same 1 vCPU class. What you pay, and what you get with it.
| Uno Small | Budget VPS Hetzner, OVH, Contabo | Cloud VPS DigitalOcean, Vultr, Linode | |
|---|---|---|---|
| Price a month | $5 | $4.5–6.5 | $5–7 |
| Root and Docker | yes | yes | yes |
| Who sets it up | an AI agent, or you | you, over SSH | you, over SSH |
| When idle | asleep, memory kept, earns boosts | runs and bills all month | runs and bills all month |
| Cloud storage | 25 GB included | separate product | separate product |
| Sites and forms | free, one message | set up a web server | set up a web server |
| AI in the plan | 5 h a month, Uno AI optional | no | no |
Prices checked 2026-10-01. Budget hosts sell memory cheaper per GB. If all you need is a lot of RAM and nothing else, they are the better deal.
The short way: ask the agent to host it
In Uno Work you describe the bot to the agent, paste the token from @BotFather into a safe field, and the agent builds and runs the bot on your computer. The setup steps are in Where to host a Telegram bot so it stays online.
Plus: $20 a month
A host for your bot, from $5 a month
A computer with root and Docker that sleeps when nothing is happening. Pick the size; move up any time.
Small
A Telegram bot, a VPN, a small backend
$5/ month
- 1 vCPU · 1 GB
- 15 GB disk
- 25 GB cloud storage
- 5 boosts a month
Plus
Agents that work with your laptop closed
$20/ month
- 4 vCPU · 4 GB
- 40 GB disk
- 100 GB cloud storage
- Uno Work in the cloud
Pro
A product in production, a second MacBook
$70/ month
- 16 vCPU · 16 GB
- 120 GB disk
- 300 GB cloud storage
- Speed of a MacBook Air M4
Max
A team and many agents at once
$200/ month
- 32 vCPU · 48 GB
- 300 GB disk
- 1,024 GB cloud storage
- 5 public IPv4
Add Uno AI to any plan: 40 h of AI work a month on Small, 100 h on Plus, unlimited on Max. Today: GLM-5.3 Flash and MiMo v2.6 Pro. Or sign in with your own Claude or ChatGPT.