The Full Picture
Here's what you're paying for after completing this curriculum:
| Service | Monthly Cost | What It Does |
|---|---|---|
| Vultr VPS | ~$24 | The server that runs 24/7 |
| Vultr Backups | ~$4 | Automatic weekly snapshots |
| Anthropic API | ~$100–1,000 | Claude's brain (varies by usage) |
| Privacy.com | $0 | Virtual payment cards |
| Bitwarden | $0 | Password manager |
| Proton Mail | $0 | Dedicated email |
| Telegram | $0 | Messaging channel |
| OpenClaw | $0 | The software itself (open source) |
| Total | ~$150–1,050/month |
The biggest variable is Anthropic API usage. Here's how to control it.
Monitor Your Anthropic Spending
- Bookmark the Anthropic Usage Dashboard — check it weekly to stay on top of your spending.
- The spending limit you set in Lesson 10 is your safety net.
- Most of your cost comes from:
- Cron jobs — each scheduled run costs tokens
- Long conversations — more context = more input tokens
- Opus usage — 15x more expensive than Haiku per token
What Does Typical Usage Actually Cost?
The "$10–50/month" range is wide because it depends entirely on how you use your bot. Here's a concrete example:
Cost Reduction Strategies
1. Route Aggressively to Haiku
If you find yourself mostly asking simple questions, make Haiku the default and use Sonnet only when you explicitly need more capability.
2. Keep Conversations Focused
Start a new conversation thread rather than continuing one that's gotten very long. Long context windows mean more tokens per message.
3. Audit Your Cron Jobs
Each cron job run costs money. A daily briefing on Sonnet might cost $0.10/run. On Opus, $0.50/run. Do the math for your schedule.
4. Use the Privacy.com Spending Limit
Your "Anthropic API" Privacy.com card has a spending limit. This is your absolute ceiling — if you hit it, the API stops working but you don't get surprised by a bill.
Vultr Cost Tips
- If you're not using your OpenClaw for a while, you can snapshot and destroy the server (stops billing), then restore later
- Vultr bills hourly, so you only pay for uptime
- Don't accumulate old snapshots — they cost storage
Budget Target: Under $100/month
With sensible model routing and moderate usage, most people land at $150–300/month for a capable personal AI assistant running 24/7.
When You're Done
- Reviewed your first month's Anthropic usage
- Privacy.com spending limits set appropriately for all services
- Understand which model tier drives your costs
- Have a plan for keeping monthly costs under your budget