AI usage, costs and limits

🎯 Goal

Know what your AI costs, and make sure it never surprises you.

🧰 What you need

• The owner login
• An AI key in use

Usage

AI & Knowledge → Usage & limits (owner only) shows, for the last 7, 30 or 90 days:

• Calls, tokens and Cost, by provider and model and by use and by bot: replies, suggestions, summaries, translation, knowledge indexing and knowledge search.
• Estimated cost per day.
• Failures with reasons (key rejected, out of credit, rate limited, provider down). Every failed call handed the chat to a person.
• Key activity: who added, changed, tested or removed keys and limits.

Costs are estimates from token counts and list prices. Your provider’s invoice is the real figure. You can override the price per model; blank uses the provider’s list price where known.

Limits

• Monthly spend cap (USD, estimated): leave empty for no cap. The box shows this month’s estimated spend.
• AI replies per conversation, per 24 hours: leave empty for unlimited. Stops one chatty customer (or a bot loop) from running up your bill.

When a limit is hit, AI stops and the chat goes to your team. You get an alert.

login.meapi.in/ai?tab=usage

• Limits: Monthly spend cap (USD, estimated) with this month's spend so far, and AI replies per conversation, per 24 hours. Empty = no cap / unlimited.
• Estimated cost per day for the chosen period.
• Tables By provider and model and By use and by bot: tokens, calls and cost.
• Failures with their reasons, and Key activity: who added, changed, tested or removed keys and limits.

When something goes wrong

If a key stops working or the cap is reached, AI replies pause and chats go to people. The owner gets an alert in the panel (AI replies paused: … with a Fix button) and by email. Fix the key or raise the cap, and AI resumes immediately.

Tip

Set the budget at your provider too (OpenAI Limits, Google Cloud budgets). Two safety nets are better than one.

What’s next

AI providers →Change or add keys.