Tracking Usage and Token Costs
AI Hub meters every operation in Liferay Tokens (LRT), a single unit that folds the underlying model, content-safety, and embedding costs into one figure you can track and budget against. You can check your remaining balance on the Activity page, and AI Hub enforces a spending limit so an account cannot exceed its purchased capacity.
AI Hub converts every operation into LRT:
- Language model calls — the requests an agent sends to the model and the responses it receives.
- Agentic execution steps — each language-model or AI-decision step an agent runs. An agent with two language-model steps counts as two execution steps per call.
- Content-safety checks — the screening that guardrails apply to inputs and outputs.
- Text embeddings — created when AI Hub indexes your content for search.
LRT are purchased in blocks, and consumption is determined using a fixed set of conversion rates. The conversion rate for each operation is locked to each block when you purchase it. Liferay can change the rates, but a change applies only to blocks purchased on or after the new rate takes effect — blocks you already own keep their original rates.
Prepaid Token Blocks
AI Hub consumption is paid from prepaid token blocks purchased through your Liferay Marketplace account. Each block has a set size in Liferay Tokens (LRT), records a purchase date, and locks in the conversion rate version that was in effect when you purchased it. AI Hub draws down sequentially: your oldest block is consumed before newer ones. Each operation’s raw token usage is converted to LRT at the rate locked to the block being drawn down, then subtracted from that block’s remaining balance; when a block can’t cover the full cost, the remainder spills to the next block.
Your account’s remaining balance appears on the Activity page.
Viewing Your Token Balance
AI Hub summarizes your account on the Activity page, including how much prepaid balance remains.
-
In the AI Hub left navigation, click Activity.
-
Review the summary cards:
- Agents: the number of agents in your account
- Chatbots: the number of chatbots in your account
- Remaining Balance Liferay Tokens: the LRT still available across your prepaid blocks
The balance updates as agent runs complete, so it reflects your current consumption.
When the remaining balance reaches zero, AI Hub blocks further agent runs until you add prepaid balance.
Spending Limits
AI Hub enforces a hard limit on consumption: an account can run agents only while it has prepaid balance remaining. There is no separate spending cap to configure — the purchased balance is the limit. When every prepaid block is exhausted, AI Hub blocks further agent runs for that account rather than throttling them, and the chatbot responds with You have exceeded your quota instead of an answer. To resume, you must purchase a new block from your Liferay Marketplace account.
Concurrent Request Limits
Beyond the spending limit, AI Hub caps how many agent requests an account can run at the same time. Each account can run up to ten concurrent requests. When an account is already at the cap, AI Hub rejects the new request with the message You have exceeded your concurrent request limit, so the calling application can wait and retry. A request slot is held only briefly: if a request stalls or is abandoned, AI Hub reclaims its slot after one minute, with no administrator action required.