Most Discord bot hosting decisions start with one question: how much RAM do I need? The honest answer is: it depends on your library, cache settings, number of servers, database usage, and features.
This guide gives practical starting points for 2026.
Quick RAM table
| Bot type | Good starting RAM | Notes |
|---|---|---|
| Learning or ping bot | 64-128 MB | Minimal dependencies |
| Small moderation bot | 256 MB | Basic commands and logs |
| Reaction role bot | 256-512 MB | Depends on event volume |
| Ticket bot | 512 MB | Transcripts and database usage add memory |
| Economy bot | 512 MB-1 GB | Cache and database clients matter |
| Music bot | 1 GB+ | Lavalink or audio processing increases needs |
| AI bot | 1 GB+ | Queues and API context can grow quickly |
| Multi-server public bot | 1 GB+ | Sharding may be needed later |
Start small, then upgrade when monitoring shows pressure.
Why RAM usage grows
A bot's memory usage comes from more than your code file. It includes:
- Node.js or Python runtime
- Discord library cache
- Command registry
- Database client
- HTTP clients
- Logging buffers
- Scheduled jobs
- Loaded modules and plugins
A bot that looks tiny in source code can still use hundreds of MB once dependencies are loaded.
Discord.js memory notes
Discord.js is powerful and friendly, but caching can become expensive. Guilds, channels, roles, members, messages, and reactions may all use memory depending on intents and cache settings.
Ways to reduce memory:
- Only enable intents you need
- Avoid caching unnecessary members
- Use sweepers for old cache entries
- Store persistent state in a database, not in memory
- Avoid loading large JSON files at startup
discord.py memory notes
Python bots can be very efficient, but memory still grows with cogs, background tasks, HTTP sessions, and caches.
Ways to reduce memory:
- Close HTTP sessions properly
- Avoid global lists that grow forever
- Use async database drivers carefully
- Log errors without storing huge objects
- Restart gracefully after deploys
Signs your bot needs more RAM
Watch for:
- Random exits without clear code errors
- Out-of-memory messages in logs
- Slow command responses during busy times
- Crashes after joining more servers
- Crashes during large moderation or ticket tasks
- Increasing memory use that never drops
If memory grows forever, also check for leaks. Upgrading RAM hides the symptom but does not always fix the cause.
How to size by server count
Server count is only one part of the story. A bot in 20 huge servers can use more resources than a bot in 200 tiny servers.
Better signals:
- Commands per minute
- Events per minute
- Number of cached members
- Ticket transcripts per day
- Database writes per minute
- External API calls per command
Plan recommendation
Use the free plan for tests and very small bots. Use 512 MB for active private bots, moderation helpers, ticket bots, and small community tools. Use 1 GB or more for public bots, music workloads, AI integrations, economy systems, or anything with multiple background jobs.
Final rule
Choose the smallest plan that stays stable during your busiest hour. Monitor logs, upgrade when usage is real, and keep your code efficient.