AI Credits & Usage

How many hosted-AI credits each plan adds per seat, what one AI request costs, and where an administrator sees what is left.

Credits per seat

Every plan adds AI credits to your workspace each month for every seat. The workspace pool is credits per seat × members. It resets on the 1st of each month (UTC) and unused credits do not roll over.

Plan	AI credits per seat / month
Basic	500
Standard	1,500
Pro	5,000
Enterprise	Unlimited (custom contract)

For example, a Standard workspace with 10 members has a pool of 15,000 credits per month.

What one request costs

A request costs credits by how heavy the model is — not by how many tokens it uses — so a cheap model and a premium model are easy to compare.

Model type	Credits per request
Fast / low-cost models, Semantic Search embeddings and rerank	1
Standard chat models	5
Premium / reasoning models	10

Every AI feature draws from the same pool: every AI Assistant panel, the writer tools, the agents and similarity search. Only a request that succeeds uses credits; a refused or failed request is not charged.

Hosted AI only — local models are free

Credits are used only by hosted AI

Hosted AI means a cloud model called through TotalApp (for example the Cloud LLM engine). Local models never use credits: Ollama, Local LLM (llama.cpp) and the in-browser models (WebGPU) run on your own machine or network.

See what is left — Admin → Usage

Administrators open Admin → Usage to see the credits remaining this month, the credits used today and this month, the plan and its monthly pool, and a daily usage chart for the last 7, 30 or 90 days. When the pool is almost used up, a warning is shown. Once it is used up, new hosted-AI requests are refused until the monthly reset or a plan upgrade.

Extra limits

Inside the plan pool, administrators can add their own limits under Settings → Agentic → AI Credits: a per-request limit (blocks models that cost more than the limit), a daily limit and a monthly budget. See Agentic Settings.

FAQ

Do I use credits when I run a local model?

No. Ollama, Local LLM and in-browser models never use credits.

What happens when the credits run out?

New hosted-AI requests are refused with a message that the plan's monthly credits are used up. Local models keep working. The pool refills on the 1st of the month (UTC), or you can upgrade the plan.

Do I get more credits when I add members?

Yes. The pool is credits per seat times the number of members, so it grows with your team.