Usage Limits & Cost Control

Stay in control of API spending with per-user limits, global caps, and usage monitoring.

Why Set Limits?

Control API costs — Prevent unexpected bills from your AI provider
Prevent abuse — Stop a single user from consuming all resources
Fair distribution — Ensure all visitors get access
System performance — Avoid overwhelming your server

Types of Limits

Message Limits

Per-User Daily Limit

Maximum number of messages a single user can send per day. Resets every 24 hours. Identified by IP address.

Global Daily Limit

Total messages for all users combined per day. Protects against unexpected spikes and keeps you within a fixed API budget.

Budget Limits (Token & Cost)

In addition to message counts, you can set hard limits based on actual token consumption and dollar cost. These are ideal for keeping your API spending predictable.

Daily Token Limit

Maximum total tokens (input + output combined) consumed across all conversations in a single day. Once reached, no new messages are processed until the next day.

Daily USD Limit

Maximum dollar amount (based on calculated cost per conversation) allowed per day. Useful when you want to set a fixed daily budget in real money terms.

Monthly Token Limit

Same as the daily token limit, but accumulated over the entire calendar month. Resets on the first day of each month.

Monthly USD Limit

Maximum dollar spend per calendar month. This is the most straightforward way to cap your total costs.

💡 Combine both types Message limits and budget limits work independently. You can set both — for example, 200 messages per day and $5/month — and whichever limit is reached first will take effect.

Configuration

Message Limits

Go to AxiaChat AI → Settings → Usage & Moderation.
In the Usage (Limits) card, select the Message Limits tab.
Set:
- Messages per user per day
- Global messages per day
- Behavior when limits are reached (disable input or hide widget)

Budget Limits

Go to AxiaChat AI → Settings → Usage & Moderation.
In the Usage (Limits) card, select the Budget Limits tab (shown by default).
Set any combination of:
- Daily token limit
- Daily USD limit
- Monthly token limit
- Monthly USD limit
Choose the behavior when the budget is exceeded.

Leave any field at 0 to disable that specific limit.

When Limits Are Reached

Message Limits

Option	Behavior
Disable Input	Text field is disabled, explanatory message shown, widget remains visible
Hide Widget	Chat disappears completely, reappears when limits reset

Budget Limits

Option	Behavior
Hide widget completely	The chat widget is not rendered at all — visitors won't see anything. This is checked server-side when the page loads, so the widget never flickers into view.
Show WhatsApp button only	The chat input and messages are hidden, but the WhatsApp CTA button remains visible so visitors can still reach you through WhatsApp. Requires WhatsApp to be configured in the bot settings. If WhatsApp is not configured, the widget hides completely as a fallback.

💡 Pre-render check Budget limits are checked at page load time (not just when a message is sent). If the limit has already been exceeded, new visitors won't see the widget at all — preventing confusion and unnecessary page weight. Admins always see the widget for testing purposes.

IP Storage

To enforce per-user limits, visitor identification is needed:

IP is stored in binary format (VARBINARY) for efficiency
Only stored when per-IP limits are active
You can disable IP storage if you don't use per-user limits
Compatible with GDPR requirements (see Privacy & GDPR)

Usage Monitoring

Conversation Logs

At AxiaChat AI → Logs you can see:

Messages per day
Usage by bot
Usage by user/session

Cost Metrics

The plugin automatically calculates:

Input and output tokens per message
Estimated cost per conversation (based on model used)
Daily and monthly totals
Monthly summary with a month picker — see total tokens, cost, and per-model breakdown for any month

Provider Pricing Overview

Provider	Budget Option	Mid-Range	Premium
OpenAI	GPT-5 Nano	GPT-4o-mini	GPT-5.2
Claude	Haiku 3.5	Sonnet 4.5	Opus 4.1
Gemini	Flash-Lite (free!)	Flash 2.5	Pro 2.5

Typical estimate for an average e-commerce site (100–500 conversations/month): $5–$20/month.

Tips to Optimize Costs

Choose the right model

"Mini" and "Nano" models are significantly cheaper
Reserve premium models for complex use cases
GPT-4o-mini or Gemini Flash offer great quality-to-price ratio

Optimize context

Index only relevant content (skip privacy policy, legal boilerplate)
Limit the Knowledge Sources slider (start with 5 chunks)
Use "Current Page" mode when full embeddings aren't needed

Adjust max tokens

Reduce max_tokens if responses are too long
The plugin suggests recommended values per model

Set reasonable limits

Start with conservative per-user limits
Adjust based on actual usage patterns
Monitor regularly through the logs

💡 Gemini free tier Google Gemini offers a free API tier with generous limits. If you're budget-conscious, start with Gemini Flash-Lite for your FAQ bot — zero cost.