How AI usage works

Your subscription includes an AI budget

Every Technovault subscription (including the 30-day trial) includes amonthly AI allowance. Ask questions about your books; each answer consumes a little of it. Your exact current allowance and usage are always shown on your Billing page. When it's used up, the AI pauses until next month — it stops, it never charges you. Want more? AI plans start at $5/month and add to your included allowance any time.

The math, in one sentence

AI usage is measured in tokens (a token is roughly three-quarters of a word). Your question and the ledger data the AI reads are input tokens; its answer is output tokens. We metertokens × the published rate for your chosen engineand subtract the resulting dollars from your budget. The full rate schedule is below — what you see is what you're charged, nothing hidden.

Worked example: a typical "find my reconciliation error" question reads ~30,000 tokens of ledger data and writes a ~1,500-token answer. On a mid-priced engine that's roughly 30k × $5/M + 1.5k × $25/M ≈ $0.19. On the cheapest engine, the same question costs about a penny.

Current rates, by engine

You choose the engine per question, right in the chat. Labels show where your data is processed — that's your call to make, not ours. Your company's administrator can disable any engine for the whole company.

Unused budget expires

Budgets reset monthly; unused amounts don't roll over or refund. Cancel any paid AI plan whenever you like — you keep the remainder of the month you paid for, and you simply aren't renewed.