Capabilities: The Platform

AI cost management

AI introduces a new category of cost into every business, and it's already fiendishly complex to manage. Subscriptions, token pricing, throttling, overage charges: every organisation needs to understand what it's spending and whether the return justifies it. VibeMagic makes that possible by routing all AI calls through a central portal, logging every one, and giving decision-makers a clear view of who is using what and what it costs.

Central AI portal API
OpenRouter integration
Per-user and per-app reporting
CFO-ready dashboards

The Pricing Problem

AI pricing is already complicated, and getting more so.

There are currently two ways organisations pay for AI. Both have significant hidden complexity that makes budgeting and oversight difficult without the right infrastructure.

Subscriptions

Difficult to budget and oversee

A monthly fee (typically £20 to £200 per user) that in theory covers an unspecified amount of AI use. In practice, the economics are more complicated. High-demand users (software developers, analysts) consume far more than low-demand users, making the flat-fee model unstable. AI companies manage this through throttling, queuing, and "overage" charges that make budgeting unreliable. Enterprise or centralised billing often comes at a premium on top. Subscriptions are currently VC-subsidised and considerably cheaper than API pricing, but this gap is likely to narrow as funding pressure grows.

Token (API) pricing

Precise but hard to compare across models

Customers pay per "token", a unit of processing roughly equivalent to half an English word. In principle this is pay-as-you-go; in practice, headline rates are misleading. Different models convert input to tokens differently (up to 1.5× variation), and models also differ significantly in thinking efficiency for the same task. The only meaningful way to compare AI pricing is to run the same task across multiple models and measure the actual cost, and this needs to be done per task type, not in the abstract. Currently, API pricing is generally more expensive than subscriptions, though the gap is expected to close.

The VibeMagic Approach

Every AI call logged, costed, and attributed.

VibeMagic provides a centralised AI prompt API inside the core portal. Every app or tool built on the platform is required to use it, which means every AI call passes through a single point where it can be logged, attributed, and costed.

Central Portal API

All AI calls route through one place

Any internal app or tool that needs to use AI makes its call through VibeMagic's Portal LLM API, not directly to an AI provider. This is a requirement for all VM apps and tools, on both VM Desktop and the web platform.

The Portal API records the calling user, the app or tool, the module within the app, and optionally the project the work is associated with. These attributes travel with every call, making every piece of AI spend fully traceable before a single pound is billed.

Built-in attribution. Logging isn't an afterthought added later. It's baked into the API contract. Every VM app must pass user, app and module context; the platform enforces this, so there are no gaps in the data.

OpenRouter Integration

One interface to hundreds of models, with costs returned

The Portal API routes calls through OpenRouter, which provides a single interface to hundreds of AI models from different providers. This means VibeMagic is not locked to any single AI company. Apps can be configured to use the most appropriate and cost-effective model for each task.

Critically, OpenRouter returns the financial cost of every API call. This cost data is captured alongside the attribution information and stored in VibeMagic's central database, giving a complete, per-call cost record for every piece of AI work done across the organisation.

Model flexibility built in. Because all calls go through OpenRouter, switching a tool to a cheaper model (or a more capable one) is a configuration change, not a rebuild.

Cost Reporting

A clear picture of AI spend across the organisation

VibeMagic's reporting system gives finance and management the view they need to understand AI spend. Total spend per user shows which individuals are the heaviest AI users, useful both for cost management and for identifying who is getting the most value. Total spend per app or tool shows which capabilities are being used and whether the investment in building them is justified.

This is the data a CFO actually needs: not just a total bill, but a breakdown that connects cost to specific activities and people, enabling proper cost-benefit analysis of AI investment across the organisation.

Current Limitations

What we track, and what we're still working on.

VibeMagic's cost management approach is comprehensive for API-based AI calls within the platform. Two areas are still in development, and we're happy to discuss them with any customer for whom they're a priority.

External AI tools

VibeMagic's logging covers API calls made through the platform. It does not currently capture spend in external AI tools, such as ChatGPT or Copilot subscriptions that individual users hold independently. Tracking that spend requires a different approach, and it's something we're actively looking into.

Subscription cost tracking

Because VibeMagic apps use the API rather than individual subscriptions, the platform doesn't capture subscription spend automatically. API calls are currently more expensive than subscription rates, so VibeMagic aims to reduce the gap by routing to lower-cost models where quality allows. VM Desktop uses the user's own subscription for many of its operations.

Key Facts

What this means in practice.

Every call attributed

User, app, module and optionally project are logged with every AI call: no gaps, no anonymous spend.

Real cost data, per call

OpenRouter returns the actual cost of each API call. VibeMagic stores it alongside the attribution data: pound-by-pound visibility.

Model-agnostic routing

Calls can be routed to the most cost-effective model for each task. Switching is a configuration change, not a rebuild.

CFO-ready reporting

Spend breakdowns by user and by tool give finance the data to assess whether AI investment is delivering return.

Want to understand your AI spend?

Talk to the team about how VibeMagic's cost management tools work in practice, and how they'd apply to your organisation's AI usage.

Talk to the Team Explore the Platform