CreatorFlow · Usage Billing & Cost Architecture

Unit economics wireframe

Six measurements form the operating spine.

01Source MinutesCustomer content duration. Stable customer-facing allowance.

02Processing MinutesActual compute runtime across transcription and downstream jobs.

03AI / API UnitsExternal model tokens, audio units, searches or enrichment calls.

04StoragePrivate source + derivatives measured in GB-month.

05EgressDownloads/streaming/network transfer when material.

06Service LayerSecurity, monitoring, retries, support, QA and publishing value.

Billing rule: do not bill raw MP4 size as a primary value metric. Bill source duration and transformation/service tier; meter infrastructure-heavy behavior underneath and expose overages only where they are meaningful and understandable.

Why cloud APIs still count as competition

The customer can buy primitives—but then must build everything around them.

Cloud API path

Cheap unit → expensive assembly

A customer can call AWS, Google, OpenAI, AssemblyAI or similar services directly. But they still need secure upload/storage, job orchestration, retries, tenant isolation, deletion policy, dashboards, transformations, publishing, billing, and operations.

DIY agent path

Flexible → fragmented

A solopreneur can combine ChatGPT/agents, scripts, storage and site generation. CreatorFlow competes by making that architecture repeatable, governed, observable and hands-off.

Human path

Judgment → labor cost

A freelancer or small shop can deliver excellent results manually. CreatorFlow competes on continuity, throughput, automation and cumulative Digital Library value while retaining human-review layers where they matter.

Planning calculator

Explore margin before publishing a price.

This is a planning wireframe using internal assumptions—not a quote engine. Provider pricing and your actual hardware/cloud runtime should feed the production cost ledger later.

Source duration90 min

Processing/runtime multiplier1.5×

Private storage3 GB

Premium AI/API allocation$1.00

Target customer charge$69

Processing runtime135 min

Estimated variable cost$2.38

Gross contribution$66.62

Modeled gross margin96.6%

The low machine-cost result is intentional: raw processing can be inexpensive. Your real product price must also fund engineering, product development, security, monitoring, support, payment fees, failed/retried jobs, customer acquisition and profit.

Stripe meter wireframe

Subscription value on top. Metered consumption underneath.

Meter	Why track it	Customer visibility	Stripe aggregation
source_minutes	Primary allowance / plan consumption	Visible	Sum
processing_minutes	Protect against unusually expensive compute workloads	Overage/unlimited tiers	Sum
storage_gb_month	Persistent private-media and derivative retention	Studio/high-throughput	Sum
egress_gb	Large downloads or future streaming	Only when material	Sum
premium_ai_units	Pass-through / marked-up external AI features	Optional premium usage	Sum

Recommended customer experience: ordinary customers see a predictable monthly plan and generous included usage. Usage meters protect the backend. Only high-throughput, unlimited, or unusually expensive jobs surface detailed usage charges.