Shared AI usage

Document Studio AI writing tools, PDF Intelligence summarisation and chat, and Design Studio image generation all sit under the same commercial umbrella. You choose a plan for model tier access and allowance, then work across studios without opening a second AI vendor account for brand assets.

That matters for real packs. A proposal might use AI to draft narrative, PDF Intelligence to pull citations from a filing, and Design Studio for a letterhead — then export everything. Shared usage keeps the AI portion coherent while free editing stitches the pack together.

Roadmap studios may join the marketplace later; the platform principle remains: metered AI where models run, free human craft everywhere else. Shared usage is the economic expression of that principle.

Because the pool is shared, finance can ask a single question — how many tokens did the organisation consume this month, and in which apps — rather than reconciling three invoices with three definitions of “credit.”

Tokens are consumed when an AI tool completes successfully: generate, rewrite, summarise, translate, logo generation, and similar model-backed actions. The audit trail records app, tool, model, tokens, and outcome so the number is inspectable.

Failed calls use zero tokens. If a provider errors or a run does not complete successfully, your pool should not shrink for nothing. That policy is part of transparent usage — not a footnote.

Editing and export are free by design. Collaboration features like comments, live presence, versions, approvals, and share links are process surfaces, not generation. They do not require a calculator because they are not metered as AI usage.

There are no pre-run cost estimates that guess tokens before you click. CasperWasp shows what was consumed after the fact. Teams that need predictability learn from the ledger patterns of their own workloads rather than from a generic estimator.

Only owners and admins change plans or purchase token packs. Editors and viewers do the creative and review work without wallet authority. That separation prevents surprise top-ups while keeping drafting unblocked.

When a launch month spikes design generation and long-document rewriting, admins can top up rather than freezing the team. Shared usage makes the spike visible in one place — then solvable in one place.

Healthy norms beat anxious metering. Save versions before large AI rewrites. Prefer targeted rewrite tools over regenerating entire documents. Use stronger model tiers when the task warrants them, not by default for every grammar fix.

Pair shared usage with organisations-and-roles and the per-call audit trail. Roles decide who can spend; the ledger explains what was spent in tokens; shared usage defines that Document Studio and Design Studio draw from the same well.

Pricing communicates plan allowance and model tier unlocks. The AI models catalogue shows which text and image models each tier can reach. Shared usage is the runtime behaviour that connects those pages to everyday work.

You do not need a separate product for “AI writing credits” versus “AI design credits.” The marketplace subscription is the product; studios are surfaces; tokens are the metered unit for model calls.

INR pricing and plan details live on the pricing page. This feature page explains the usage model so buyers understand what they are joining — an AI marketplace with transparent token consumption, not a black-box seat that hides AI cost until renewal.

If you are evaluating CasperWasp against stitched point tools, shared usage is often the commercial unlock: one team, one library, one token pool, documents + design.