IFRS AI Advisory SuiteAdvisory workspace
Sign in

Token calculator

How many tokens does your filing season need?

~30k tokens per consultation
Tokens needed
3.6M
Provider cost at $0.30/1M
$1.08
Recommended plan
Professional

Plan price

$49/month

Cost per consultation

$0.41

At 98% gross margin this plan covers the provider cost of $1.08 for the month. Running on your own key on Enterprise removes the token meter entirely.

Request the Professional plan

Transparent pricing

Plans for practitioners, firms and students

One token unit across the product: roughly 30,000 tokens per full consultation. No API mark-up — or connect your own provider key and pay the provider directly.

Starter

$0/ forever

For individual accountants and students testing the engine on real questions.

  • 1,000,000 tokens once (~33 consultations)
  • Balance validator on every journal entry
  • IFRS, IAS and IFRIC citations
  • Memo (.docx) and schedule (.xlsx) export, marked as a starter draft
  • Platform-managed model only
  • Bring your own key
  • Team seats
  • Private knowledge ingestion
Claim free quota
Most popular

Professional

$49/ month

For sole practitioners and senior auditors with recurring filing deadlines.

  • 10,000,000 tokens per month (~330 consultations)
  • 1 seat included
  • Unlimited memos, schedules and disclosure notes
  • Exports without the starter mark
  • Bring your own DeepInfra, OpenAI or Gemini key
  • Priority routing on the managed model
  • Team seats
  • Private knowledge ingestion
Request upgrade

Enterprise

$149/ month

For audit firms and group finance teams with several reviewers on one file.

  • 5 team seats included, extra seats $25/month each
  • Unlimited usage on your own provider keys
  • Private firm knowledge ingestion (manuals, past memos)
  • White-label memo footer and firm letterhead
  • Audit trail export for the working-paper file
  • Dedicated container and priority support
Request consultation

Seats beyond the 5 included on Enterprise are $25/month each. There is no card form in the product yet: a plan request opens a pre-filled message to advisory@vacsolution.com.

Frequently asked

Clear answers for partners and controllers

How does the engine prevent unbalanced journal entries?

Every journal entry the model proposes is parsed and re-added before the answer is shown. If debits and credits disagree by more than half a cent, the response is rejected and the model is asked to correct it — the same check runs in src/core/accounting_validator.py.

How are token quotas counted?

One average consultation — prompt, retrieval context, reasoning and answer — uses roughly 30,000 tokens, so the 1,000,000 starter quota is about 30 full consultations and the 10,000,000 monthly quota about 330. The quota chip in the header shows the live balance, and every answer reports the tokens it consumed.

Can we use our own provider key instead of your quota?

Yes, from the Professional plan upward. Keys for DeepInfra, OpenAI/Azure and Gemini are stored encrypted and never returned to the browser; the settings page shows only a masked preview. On Enterprise you can route all traffic to your own account and leave the platform quota untouched.

Which standards and interpretations are covered?

The knowledge base covers IFRS 1 to 18, IAS 1 to 41 and IFRIC agenda decisions, and every answer cites the paragraph it relies on. Superseded guidance is filtered out so the model cannot quote a withdrawn treatment as current.

Is our client data used to train the model?

No. Uploaded trial balances, schedules and conversation text are stored in the firm workspace and sent to the configured model provider only to answer the question. Enterprise deployments can pin the traffic to a provider account the firm controls.

Can the exports be filed as audit documentation?

The memo follows the usual structure — objective, guidance, analysis, journal entries, preparer and reviewer sign-off — so it can be filed alongside the working papers. The reviewer sign-off remains a human signature: the tool drafts, the partner approves.