Outcome-based billing infrastructure for AI agents that evaluates event signals over time, settles verified outcomes and carries auditable context into billing.
Product sources: https://www.thewitn.com/ ; https://www.thewitn.com/docs/introduction ; https://www.thewitn.com/docs/guides/how-witn-works
witn is billing infrastructure for AI-agent builders that measures whether a defined outcome actually became billable instead of charging for a proxy such as seats, tokens or API calls.1 Its model is event-driven: a task defines the billable condition and settlement window, a customer carries a rate card, and an outcome changes state as new signals arrive.2
The API is versioned under `/v1`. The quickstart shows asynchronous `202 Accepted` creation and event submission, with later signals able to move an outcome from provisionally satisfied to failed before settlement.3 The default settlement window is 72 hours and can be configured up to 180 days.2 The public product also describes simulation/replay and an ordered decision trail intended to explain invoice charges.1
witn is currently demo-led; no public self-serve numeric subscription price was verified. Its rate-card model lets a customer be assigned task-specific `price_per_unit` values, and the price is snapshotted when an outcome opens so later rate-card edits do not alter historical outcomes.4
That architecture is flexible for outcome pricing, but the absence of public platform pricing makes value-for-money comparison difficult. Buyers need to validate commercial terms directly and compare them with both usage-billing vendors and an internal implementation.
The current product and API surface are documented in useful technical detail: outcome states, billable conditions, settlement, event handling and rate cards are all explicit.23 The homepage says each signal and decision is recorded in order to support charge disputes.1
Independent validation is the main evidence gap. The canonical 50-source panel returned no eligible exact-product external numeric observation; launch-directory/showcase mirrors and unrelated products sharing the name were excluded rather than treated as reviews. This rating is therefore desk-evidence-led. No hands-on production reliability, dispute handling or billing-accuracy test is claimed.
The strongest fit is an AI-agent vendor whose customers care about completed results rather than activity metrics: support resolution, coding work, SRE actions, voice-agent outcomes or other events that can be emitted into an API.1 It is particularly relevant when reversals matter because an apparently successful event can later be reopened, escalated or invalidated before settlement.2
It is less compelling for teams satisfied with ordinary usage metering, for buyers that require published pricing before evaluation, or for enterprises that need extensive independent reliability and compliance evidence before adopting a billing-critical dependency.
clear product focus; stateful settlement rather than one-shot event metering; explicit reversal handling; reusable customer rate cards; a documented REST API; asynchronous event flow; simulation/replay; and compatibility with existing event/billing stacks including Stripe, Segment, RudderStack and PostHog.13
no verified public numeric platform pricing; very limited independent exact-product review or user-rating coverage; public evidence is dominated by the vendor's own documentation; and production-scale reliability, security/compliance depth and real-world billing accuracy were not independently validated.
Products matched by shared tasks, audience, workflow, product type and capabilities.