Cost control and rate limits
Day 15 of 30 · Generative AI 2026: Build AI Apps and Agents
One-liner: Add guardrails that keep AI costs predictable.
Time: 20 to 30 min
Deliverable: Cost Guard Checklist and Limits
Learning goal
You will be able to: Set cost and rate limits that protect your AI product.
Success criteria (observable)
- A cost guard checklist is complete.
- Rate limits are defined for key endpoints.
- A monthly cost cap is written.
Output you will produce
- Deliverable: Cost Guard Checklist and Limits
- Format: Checklist plus limits table
- Where saved: Course folder under
/generative-ai-2026-build-ai-apps-and-agents/
Who
Primary persona: Digital nomad managing AI costs Secondary persona(s): Paying users who expect stable service Stakeholders (optional): Collaborators
What
What it is
A small set of rules that limit usage, protect your budget, and keep the service stable. It combines a cost cap with rate limits so one user cannot burn the budget.
What it is not
It is not a full finance system or a replacement for pricing strategy. It is a safety layer for early stage products.
2-minute theory
- AI costs can grow quickly with heavy use.
- Rate limits protect both stability and budget.
- Clear caps make it safe to test pricing.
Key terms
- Rate limit: A rule that limits requests per user or per time window.
- Cost cap: A maximum amount you are willing to spend per period.
Where
Applies in
- API routes
- Billing logic
Does not apply in
- One off manual tests
Touchpoints
- Usage logs
- Billing dashboard
- Alerting rules
When
Use it when
- You open the app to real users
- You start charging money
Frequency
Set once, adjust as usage grows
Late signals
- Unusual spikes in usage
- Bills higher than expected
Why it matters
Practical benefits
- Predictable costs
- Fewer outages
- Better control of free usage
Risks of ignoring
- Surprise bills
- Abuse by heavy users
Expectations
- Improves: cost stability and safety
- Does not guarantee: perfect margins
How
Step-by-step method
- Choose a monthly cost cap.
- Define per user rate limits.
- Add a daily limit for free users.
- Set alerts for spikes.
- Review usage weekly.
Do and don't
Do
- Start with conservative limits
- Log usage per user
Don't
- Offer unlimited free usage without controls
- Wait for a bill to react
Common mistakes and fixes
- Mistake: No cap. Fix: Set a small monthly cap.
- Mistake: No per user limits. Fix: Add rate limits per user.
Done when
- A monthly cost cap is written.
- Rate limits are defined.
- Alerts are configured.
Guided exercise (10 to 15 min)
Inputs
- Expected usage per user
- Current pricing hypothesis
Steps
- Estimate cost per request.
- Set a monthly budget.
- Define per user limits.
Output format
| Field | Value |
|---|---|
| Cost per request | |
| Monthly cap | |
| Rate limits | |
| Alert trigger |
Pro tip: Start with a low cap and increase after you see real usage.
Independent exercise (5 to 10 min)
Task
Create a rule for heavy users and document the action.
Output
A heavy user rule and response.
Self-check (yes/no)
- Is a monthly cap defined?
- Are per user limits defined?
- Are alerts configured?
- Is usage reviewed weekly?
Baseline metric (recommended)
- Score: 3 of 4 checks met
- Date: 2026-02-06
- Tool used: Notes app
Bibliography (sources used)
OpenAI Pricing. OpenAI. 2026-02-06. Read: https://platform.openai.com/pricing
API Rate Limiting Guide. Cloudflare. 2024-01-01. Read: https://developers.cloudflare.com/rate-limits/
Read more (optional)
- Usage Based Pricing Why: Align costs with pricing plans. Read: https://www.profitwell.com/recur/all/usage-based-pricing