Setting Usage Caps and API Keys for Client Sites
How Do You Set Usage Caps and API Keys for Individual Client Sites?
Agencies set usage caps per client to control cost predictability across a portfolio, and manage API keys — either shared under the agency’s own account or client-specific, depending on the platform’s structure — to keep each client’s usage properly separated and billable. Getting this configuration right early avoids a genuinely common agency headache later: a single high-volume client unexpectedly consuming a disproportionate share of shared resources, or unclear cost attribution when it’s time to bill clients for actual usage.
Why Usage Caps Matter for Agencies Specifically
Unlike a single business managing its own chatbot, an agency is managing cost and usage across many separate client relationships simultaneously, often with different pricing arrangements per client. Without usage caps, a single client experiencing an unusual traffic spike — a viral social post, an unexpected PR mention — could consume a disproportionate share of shared AI provider costs, an issue that’s genuinely worth preventing proactively rather than discovering after an unexpectedly large bill.
API Key Configuration Options
| Approach | How It Works | Best For |
|---|---|---|
| Shared agency-level API key | All clients draw from one centrally managed key, agency absorbs and re-bills usage cost | Agencies wanting simpler centralised billing management |
| Client-specific API keys | Each client uses their own key, either their own account or a dedicated one per client | Agencies wanting clean cost separation and direct client billing |
Setting Sensible Usage Caps
A reasonable starting cap should be based on the client’s realistic expected traffic volume, with some margin for genuine growth — not set so low that normal usage triggers warnings constantly, but not so high that a genuine anomaly goes unnoticed until costs have already accumulated significantly. Review actual usage data after the first month or two and adjust caps based on real observed patterns rather than an initial guess alone.
What Happens When a Cap Is Reached
Worth understanding clearly before it happens in practice: does the chatbot stop functioning entirely once a cap is hit, fall back to a more limited mode, or simply generate an alert for the agency to review and adjust? Each behaviour has different implications for the client’s visitor experience, so it’s worth configuring this deliberately rather than discovering the actual behaviour unexpectedly during a real traffic spike.
Billing Considerations
- Decide upfront whether usage cost is included in a flat retainer or billed separately based on actual consumption — both are reasonable models, but ambiguity here creates friction later.
- Set client expectations about what happens during unusual traffic spikes — a launch, a viral moment — so there’s no surprise if a temporary cap adjustment or additional cost applies.
- Keep clear records of per-client usage for transparent billing conversations, particularly for agencies passing through actual usage cost rather than a flat fee.
A Realistic Scenario
An agency sets a baseline monthly usage cap for each client based on their typical site traffic, with client-specific API keys for clean cost tracking and direct billing. When one client runs a successful marketing campaign that drives significantly more traffic than usual, the agency’s centralised dashboard flags the approaching cap ahead of time, giving them a chance to proactively raise it and inform the client of the associated cost increase — rather than the chatbot unexpectedly stopping mid-campaign with no warning.
The Outview AI Chatbot‘s Agency tier includes per-client usage tracking and configuration, giving agencies visibility and control across their full client portfolio.
FAQ
Should every client have the same usage cap?
No — caps should reflect each client’s realistic traffic volume and business needs, not a uniform default applied regardless of how different clients’ actual usage patterns are.
What happens if I don’t set usage caps at all?
You lose an important early-warning mechanism for unusual usage spikes, which can lead to unexpectedly high costs discovered only after they’ve already accumulated, rather than caught proactively.
Can clients see their own usage data, or is this agency-internal only?
This depends on how the agency chooses to structure client-facing reporting — some agencies share usage transparency directly with clients; others manage it entirely internally as part of a flat-fee service.
Key Takeaways
- Usage caps prevent a single client’s traffic spike from creating unexpected agency-wide cost.
- Choose between shared or client-specific API keys based on your billing and cost-tracking needs.
- Set caps based on realistic traffic expectations, and review and adjust after real usage data comes in.
- Decide and communicate upfront how usage cost fits into your client billing structure.
See per-client usage controls with Outview’s Agency tier.