From free trials to paying users: 6 billing platforms for AI products compared

The AI billing platforms AI app builders evaluate include Credyt for real-time billing on prepaid balances with per-usage authorization, Orb for invoice-based usage billing with custom SQL metrics, Lago for open-source AGPLv3 self-hosted billing, Stigg for real-time entitlements and credit orchestration over Stripe or Zuora, Flexprice for open-source hybrid pricing with first-class wallets, and Stripe Billing as the subscription-first baseline with a metering layer. These six platforms split on one axis: whether usage is authorized and billed in real time or captured by a metering layer and reconciled at cycle end. Most AI app builders pair Stripe with one of the purpose-built tools to add the primitives Stripe alone does not ship.

At a glance: ranking

Rank	Tool	Best for
1	Credyt	Real-time billing on prepaid balances with per-usage authorization for AI products on variable-cost infrastructure
2	Orb	Invoice-based usage billing with custom SQL metrics, pricing simulation, and dimensional pricing
3	Lago	Open-source AGPLv3 billing for engineering-led teams with code transparency or compliance requirements
4	Stigg	Real-time entitlements and credit orchestration on top of Stripe, Zuora, or Chargebee
5	Flexprice	Open-source hybrid pricing with a no-code dashboard and first-class wallets, for small to mid-stage teams
6	Stripe Billing	Subscription-first baseline for AI SaaS; pair with a purpose-built platform for prepaid balance and per-usage authorization

sbb-itb-b2281d3

Why AI products need more than just Stripe

Stripe is the default for any web product, and most AI app builders start there. Stripe Billing ships subscriptions, invoices, and a usage-based metering layer (Meter Events v2 API, 100,000 ops per second). For subscription-first AI SaaS with a small metered overage, that combination works.

What Stripe does not ship is the real-time control layer AI products need on top. Stripe meters usage and reconciles it into an invoice at cycle end. By the time the invoice generates, the cost is already incurred. AI app builders need three primitives Stripe alone does not provide:

Real-time spend caps, not invoice-end reconciliation. Every AI request costs real dollars. A user signing up with a throwaway email can burn through hundreds in inference before the cycle closes. Concurrent requests from the same account can drive a balance negative in seconds.

The defense is a prepaid balance the platform checks before each action. If the balance runs out, the next action is blocked. Stripe doesn't ship that primitive. Teams build it on top of usage records and webhooks, which is a multi-month engineering project most AI app builders skip by reaching for purpose-built billing.

Entitlement primitives for promo credits, not invoice adjustments. Small AI apps win first users with promo credits, free trial grants, referral bonuses, and price experiments. An image generator launches with $5 in free credits on signup. A voice tool gives 100 free seconds per month that refresh on the first.

A coding assistant bundles a monthly credit allotment with its pro plan and lets users top up when they run out. A study app gives three free sessions, then asks for a small one-time top-up. Each pattern needs entitlement primitives. Credit grants with priorities, expiry rules, drawdown order, and per-customer hard caps. Stripe treats these as invoice adjustments, not native objects. Every promo experiment becomes a code change.

Per-customer authorization at scale, not enterprise contracts. Mid-market and enterprise customers settle on invoices that their AP team pays. There is a contract, a procurement process, and recourse if a buyer overshoots. AI app builders ship at the other end of the curve. Thousands of accounts at $5 to $30 per month.

The math inverts. Fraud, failed top-ups, and concurrent-request abuse each chip at revenue at rates that compound across thousands of accounts. There is no contract to fall back on, only product-level enforcement. Per-customer authorization, per-customer caps, and real-time fraud signals matter more than the enterprise contract primitives Stripe Billing ships.

AI app builders typically pair Stripe with one of the five purpose-built platforms below, or pick a real-time platform that handles both payments and authorization end-to-end.

AI billing feature comparison

Two axes decide most evaluations. Whether authorization happens before or after usage, and whether the customer's balance is a first-class primitive or an add-on bolted onto an invoicing engine.

Feature	Credyt	Orb	Lago	Stigg	Flexprice	Stripe Billing
Architecture	Real-time, end-to-end	Invoice-based	Invoice-based	Real-time orchestration over downstream billing	Hybrid (invoice-based + wallet credit-debit)	Subscription-first with metering layer
Usage authorization	Per-usage	Post-usage	Post-usage	Per-usage entitlement; billing post-usage downstream	Post-usage	Post-usage
Customer balance	First-class wallet primitive	Add-on (credits-on-invoice)	Add-on (up to 5 wallets)	First-class wallet primitive	First-class wallet primitive (invoice-driven debit)	Add-on (credit grants on invoice)
Multi-asset support	Native (USD, tokens, GPU hours, custom)	USD-with-labels	USD-with-labels	USD-with-labels	USD-with-labels (fixed conversion)	USD only
Open source	No	No	Yes (AGPLv3 core)	No	Yes (AGPLv3 core)	No
Customer portal	Drop-in, branded, self-service top-up	Pre-authenticated, no self-service top-up	Premium-tier, pre-authenticated	Embeddable React, JavaScript, and Vue widgets	Drop-in, read-only	Hosted, invoice-centric
Auto top-up	Customer-controlled or platform-triggered	Platform-configured	Platform-configured	Platform-configured	Platform-configured	Not available
Pricing transparency	Public ($1/MAW after 10 free wallets)	Sales-led, not public	Sales-led on cloud tiers	Public ($448/mo Growth floor)	Public ($500 / $1,000 per month)	Public (0.5% / 0.8% on recurring)

Tool by tool

The six AI billing platforms below are organized by architecture and use case. Real-time platforms (Credyt end-to-end, Stigg as orchestration) authorize and debit the customer's balance as usage happens. Invoice-based and hybrid platforms (Orb, Lago, Flexprice) capture events through a metering layer and reconcile them into an invoice at cycle end.

Stripe Billing is the subscription-first baseline. Each card below covers positioning, pricing, strengths, and trade-offs in the same structure so evaluations can run in parallel.

1. Credyt

Credyt is real-time, end-to-end billing infrastructure built for AI products where infrastructure costs hit before any invoice cycle runs. The platform queries Credyt's Wallet API before allowing an action; if the customer's balance is sufficient, the action proceeds and Credyt prices and debits the wallet atomically. Multi-asset balances hold USD, tokens, GPU hours, and custom units in one customer wallet, and the drop-in branded portal includes self-service top-up out of the box.

Best for: Real-time billing for AI products with variable, per-request infrastructure costs.

Pricing: $1 per Monthly Active Wallet (MAW), with the first 10 wallets free and the first 1M events per month free. No revenue percentage. No seat fees. At 100 active customers: ~$90/month plus pass-through PSP fees.

Strengths:

• Per-usage authorization. The platform queries Credyt's Wallet API before performing the action. If the balance is sufficient, the platform proceeds and Credyt debits the wallet atomically. The decision to allow or block belongs to the platform; Credyt provides the real-time balance state and the atomic debit.
• Native multi-asset wallets. A single customer wallet can hold USD, tokens, GPU hours, and any custom unit in parallel.
• Drop-in customer portal. Branded, self-service top-up included; no frontend engineering required.
• Event-level cost attribution. Profitability is tracked per customer, per feature, and per workload, down to individual events.
• MCP server for AI coding tools. Builders add billing through a single prompt inside Cursor, Windsurf, Claude Code, Codex, Lovable, Bolt, Replit, or V0.
• Public, transparent pricing. No sales call to see the rate.

Trade-offs:

• Smaller partner ecosystem than incumbent platforms.
• Cloud-only deployment. No self-hosting option.
• No first-class enterprise contract management. Multi-year negotiated commitments with backdating and true-ups are not native.
• Requires customers to prepay into a balance. Not a fit for post-pay enterprise billing.

Learn more: Credyt's real-time billing infrastructure.

2. Orb

Orb is invoice-based usage-based billing for engineering-led teams that treat pricing as a core product function. The platform is built around custom SQL metrics, pricing simulation against historical data, and dimensional pricing. Named customers include Vercel, Replit, Supabase, Redis, Neo4j, and LaunchDarkly.

Best for: Invoice-based usage billing with custom SQL metrics, pricing simulation, and dimensional pricing.

Pricing: Not publicly available; sales-led. Orb removed its public pricing in 2024; current rates require sales contact.

Strengths:

• Custom SQL metrics. No bespoke engineering for unusual metering logic.
• Pricing simulation. Validate pricing changes against historical usage before deploying.
• High-volume event processing. Hosted rollups for millions of events per second.
• Dimensional billing. Multi-axis pricing across region, model, tier, and customer segment.
• Agentic Payment Methods. Product support for AI agents executing payments on behalf of end users, launched in 2026.

Trade-offs:

• Post-usage invoicing only. Threshold invoicing can trigger mid-cycle but remains post-consumption; there is no real-time wallet debit.
• No self-service top-up UI in the customer portal. The pre-authenticated portal shows usage and invoices, but customers cannot top up balance themselves.
• Auto top-up is platform-configured. The platform sets top-up rules on behalf of customers.
• Implementation overhead. Setting up metering, pricing models, and invoice workflows requires engineering resources and billing expertise.

Learn more: Orb.

3. Lago

Lago is an open-source AGPLv3 billing platform for usage-based, subscription, and hybrid pricing. The core is free to self-host; managed cloud and enterprise tiers are sold separately. Customers include Mistral AI, PayPal, Groq, Synthesia, Arcee AI, Daytona, and Laravel.

Best for: Open-source AGPLv3 billing for engineering-led teams with code transparency or compliance requirements.

Pricing: AGPLv3 core is free to self-host (infrastructure and engineering cost still apply). Cloud Business and Enterprise tiers both require sales contact as of April 2026. During the Series A discussion in 2024, commenters cited a starting cloud tier of $3,000/month; the company has since simplified pricing to two tiers and removed public price points.

Strengths:

• Open-source AGPLv3 core. Forkable, auditable, no license cost on the core.
• Deployment flexibility. Cloud, self-hosted Docker Compose, on-prem, VPC, or custom infrastructure.
• Extensive integration ecosystem. Stripe, Adyen, GoCardless, Salesforce, HubSpot, NetSuite, Xero, cloud marketplaces.
• SOC 2 Type II certified.
• Lago AI agents and MCP server, both shipped in early 2026, perform operational billing actions beyond Q&A.

Trade-offs:

• Invoice-based architecture. The authoritative wallet balance updates at invoice finalization, not on event ingestion.
• No pre-usage authorization primitive. Observe-only model; the platform cannot query the wallet and block a specific action before it runs.
• Material premium gating. Customer portal, credit notes and refunds, automatic dunning, tax integrations, CRM/CPQ/accounting connectors are paid-tier.
• Self-hosted production deployments require Postgres, Redis, multiple Lago services, observability, and ongoing engineering for upgrades.

Learn more: Lago.

4. Stigg

Stigg is a real-time monetization control layer that sits between a product and its billing system. Teams use a no-code console to define entitlements, pricing, and customer-facing widgets. Stigg orchestrates these definitions downstream to Stripe, Zuora, or Chargebee, where payment collection actually happens. Customers include Webflow, Miro, AI21, Cloudinary, New Relic, PagerDuty, and Upwork.

Best for: Real-time entitlements and credit orchestration on top of Stripe, Zuora, or Chargebee.

Pricing: Sandbox free; Growth $448/month minimum, billed annually with a 20% discount, plus $0.32 per self-service subscription, $32/seat, and $0.55/GB of events; Scale custom. At 100 customers: per-unit cost computes to ~$192, but the $448 contract minimum applies.

Strengths:

• Typed entitlement primitive. getMeteredEntitlement(customerId, featureId, { requestedUsage }) returns allowed or blocked with remaining quota in a typed payload, with documented sub-100ms P95 latency.
• Unified feature model. Boolean, Numeric, and Metered features in one primitive.
• Credits suite. Stacked grants with priority order, expiry, prepaid pools, custom consumption formulas, and an append-only ledger.
• No-code pricing console. Non-engineers configure plans, entitlements, and credit rules without engineering tickets.
• Billing-system agnostic. Connects to Stripe, Zuora, Chargebee, AWS Marketplace, Apple App Store, and custom webhooks.

Trade-offs:

• No hosted redirect-style billing portal. Customer billing UI is embed-only.
• No profitability analytics. Usage and revenue reporting only; no cost ingestion for margin analysis per customer or workload.
• Integration catalog depth varies. Stripe and Zuora are the primary downstream billing integrations; Chargebee is referenced but less deeply documented, and ERP-wide accounting integrations like QuickBooks are not listed. Stigg's MCP server is available in public beta as of April 2026 .
• Growth contract minimum applies even when per-unit cost is below the floor. Production deployments described as "weeks for meaningful packaging, months for complex" in Stigg's own case-study material.

Learn more: Stigg.

5. Flexprice

Flexprice is an open-source hybrid billing platform for AI and SaaS companies running usage-based, credit-based, or hybrid pricing. The wallet primitive is first-class, the pricing dashboard is no-code, and the AGPLv3 core can be self-hosted on Docker Compose. Customers include Simplismart and Segwise.

Best for: Open-source hybrid pricing with a no-code dashboard and first-class wallets, for small to mid-stage teams running usage-and-credit models.

Pricing: Free tier (100K events/month, 3-month validity). Starter at $500/month ($400 annual, 10M events). Premium at $1,000/month ($800 annual, 25M events). Enterprise and on-prem custom. At 100 customers and $2,000 monthly revenue: $500 to $1,000/month plus pass-through PSP fees, depending on which wallet features are needed.

Strengths:

• First-class wallet primitive. Wallets are core infrastructure with full lifecycle docs: create, top-up, auto-recharge, debit, alerts.
• No-code pricing dashboard. Non-engineers configure plans, entitlements, and credit rules without engineering tickets.
• Open-source AGPLv3 core. Forkable, auditable, self-hostable on Docker Compose or AWS ECS/EKS.
• Real-time event metering. ClickHouse-backed analytics with sub-second ingestion.
• Multiple PSP integrations. Stripe, Razorpay, Paddle, and Chargebee supported out of the box for broad regional coverage.

Trade-offs:

• Usage-based wallet debit is invoice-driven, not event-driven. Charges accumulate and debit the wallet when an invoice is generated and paid, not on each event.
• Customer portal is read-only. Customers can view balances and invoices but cannot self-service top-ups.
• Auto top-up is merchant-controlled. End customers cannot set their own threshold or top-up amount.
• Wallet credits use a fixed USD conversion. No native independent non-monetary assets; tokens and custom units are mapped to a currency ratio.
• Open-source core covers metering and invoicing only. Advanced wallet features (real-time prepaid balance, recurring top-ups, entitlement management) require the paid cloud tiers .

Learn more: Flexprice.

6. Stripe Billing

Stripe Billing is Stripe's subscription and usage-based billing product, tightly integrated with Stripe's payments, tax, and revenue recognition stack. The Meter Events v2 API processes 100,000 ops per second, and Stripe added Metronome's metering and contract platform to the family on January 14, 2026. Customers include Anthropic, OpenAI, Shopify, and Figma.

Best for: Subscription-first AI products with a small metered overage. The default Stripe stack still works; layer a purpose-built platform on top when cost-per-call invariants change with every model switch.

Pricing: Starter at 0.5% on recurring payments; Scale at 0.8%; Enterprise custom. Plus Stripe Payments processing (2.9% + $0.30 per successful card charge on US cards, separate). At 100 customers and $2,000 monthly revenue: ~$10 (Starter plan) plus ~$62 in PSP fees, for ~$72/month total before adding any purpose-built billing layer.

Strengths:

• Payments infrastructure depth. 135+ currencies, 50+ payment methods, local acquiring in many markets.
• Developer ecosystem and tooling depth. Mature SDKs, libraries, and third-party integrations.
• Enterprise integrations. NetSuite, Salesforce, Stripe Data Pipeline, ASC 606 revenue recognition.
• Smart Retries for failed cards recovers ~9% more revenue compared to scheduled retries.
• Metering layer present out of the box. Meter Events v2 API at 100,000 ops per second covers most AI products at startup throughput.

Trade-offs:

• Invoice-end billing for usage. Usage accumulates and is billed at cycle close, not in real time.
• No real-time wallet debit primitive. Spend caps and prepaid balances require external tooling or custom code on top.
• Customer portal is invoice-centric. No real-time prepaid balance UI or self-service top-up.
• No native multi-asset wallet for tokens or GPU hours; credit grants operate on monetary balances only.
• Metronome integration timeline still in progress as of April 2026; specific GA dates for combined features are not public.

Learn more: Stripe Billing.

How to choose an AI billing platform

If you bill per AI inference, per token, or per generated output and need to authorize before the model runs: Credyt fits this use case directly. Per-usage authorization on a prepaid balance with multi-asset support maps directly to real-time AI cost dynamics.

If you bill per API call with custom SQL aggregation logic and need to validate pricing changes against historical data before deploying: Orb is built for this. Pricing simulation and dimensional pricing are first-class.

If you require open-source code transparency, on-premise deployment, or data residency for compliance reasons: Lago. The AGPLv3 core lets engineering-led teams self-host and audit every line; budget for the premium-tier features the typical AI app builder still needs (customer portal, credit notes, dunning).

Stigg fits when you want to keep your existing billing system but need real-time entitlements, credit orchestration, and a no-code pricing console. It lives over Stripe, Zuora, or Chargebee without replacing either.

Flexprice fits when you want open-source plus a no-code pricing dashboard plus first-class wallets at small to mid scale. Public pricing at $500 to $1,000 per month makes the total cost predictable, and Stripe, Razorpay, Paddle, and Chargebee are supported as the underlying PSP.

Stripe Billing alone still fits subscription-first products with a small metered overage, especially for teams already running Stripe end-to-end. Pair with one of the five purpose-built platforms above when AI cost economics demand real-time control, prepaid balances, or per-customer authorization.

Bottom line

The six AI billing platforms in this comparison cover the full architectural map. Real-time end-to-end (Credyt). Invoice-based with custom metrics (Orb). Open-source self-hosted (Lago). Real-time orchestration on top of existing billing (Stigg). Open-source hybrid with no-code (Flexprice). Subscription-first with a metering layer (Stripe Billing).

Naming the architectural fork is the most useful frame for AI app builders evaluating billing software for SaaS products with variable per-request costs. Pick a real-time platform if your product needs to authorize before cost is incurred or if customers expect a prepaid balance primitive. Pick an invoice-based or hybrid platform if your customers expect monthly invoices as the primary artifact and your usage data tolerates cycle-end reconciliation.

Either path is current and valid; the question is which fits the cost economics of the product you are shipping. For real-time AI billing on prepaid balances, see Credyt's approach to real-time billing for AI products.

FAQ

What is the difference between real-time billing and invoice-based usage-based billing?

Real-time billing collapses metering and billing into a single atomic operation. Usage is recorded and billed as it happens, against the customer's prepaid balance. Invoice-based usage-based billing captures events through a metering layer throughout the cycle and reconciles them into an invoice at cycle end through a separate billing engine.

Real-time platforms (Credyt end-to-end, Stigg as orchestration) can authorize before cost is incurred. Invoice-based and hybrid platforms (Orb, Lago, Flexprice) cannot.

Can Stripe Billing handle AI usage-based billing on its own?

Stripe Billing is subscription-first with a metering add-on. Its Meter Events v2 API processes 100,000 ops per second, sufficient at startup throughput levels but billed at invoice end rather than at the moment of usage. Teams looking for alternatives to Stripe Billing for AI products with per-request cost typically pair Stripe with a purpose-built platform or switch to a real-time end-to-end alternative.

What is the cheapest billing platform for AI app builders at 100-customer scale?

AI billing and payment solutions vary in price structure at 100-customer scale. Stripe Billing at ~$72/month (0.5% on $2,000 revenue plus ~$62 PSP fees) is the lowest sticker cost but does not include a real-time wallet primitive or per-usage authorization. Credyt at ~$90/month plus PSP fees adds those two capabilities for a small premium.

Stigg's Growth contract minimum is $448/month. Flexprice's Starter is $500/month ($400 annual) and Premium is $1,000/month ($800 annual). Orb and Lago Cloud do not publish pricing.

Do open-source billing platforms like Lago and Flexprice cost less than commercial cloud platforms?

The AGPLv3 cores of Lago and Flexprice have no license cost, but production deployments require Postgres, Redis (and ClickHouse in Flexprice's case), the platform services themselves, observability, and ongoing engineering for upgrades. Material features the typical AI app builder needs (customer portal, credit notes and refunds, automatic dunning, tax integrations, CRM connectors, recurring wallet top-ups) sit behind paid tiers in both cases. Total cost of ownership at production scale often lands within range of commercial alternatives once engineering time is accounted for.

Which billing platform is the fastest path to live billing for vibe coders?

Vibe coders want public pricing, drop-in customer portals, prepaid balance primitives, and integrations with the AI coding tools they already use. Among the six platforms, Credyt and Stripe Billing are the only ones with public per-customer pricing. Credyt's MCP server connects directly to Cursor, Windsurf, Claude Code, Codex, Lovable, Bolt, Replit, and V0; builders add billing through a single prompt inside the tool they already use. Stripe Billing requires building wallet logic separately, since it has no first-class wallet primitive.