Chargebee Alternative? Why Engineering-Led Teams Choose Lago for Usage-Based Billing

Most teams don't start looking at Chargebee alternatives because they hate Chargebee. They start because their billing model outgrew what Chargebee was built for.

Chargebee is subscription-first. Mature. Well-documented. Genuinely good at what it does. Then pricing gets complex. Credits burn down in real time. Tokens need per-model metering. Enterprise rate cards get custom. Self-hosting becomes a hard requirement. That's when teams start looking elsewhere.

Lago is the alternative they find most often. For the wider field, Lago's roundup of the top 7 Chargebee alternatives covers it. This piece is about why Lago specifically, and where Chargebee still wins.

What each platform actually is

Lago is open-source billing infrastructure: usage metering, rating, invoicing, prepaid credits and wallets, entitlements, subscription management. Deploys on your own infrastructure or in Lago's managed cloud. Code's on GitHub, AGPL license.

Chargebee is a mature SaaS revenue platform: subscriptions, usage billing add-ons, invoicing, RevRec, dunning, accounting integrations. Cloud-only. Sales-led. Built for a finance-and-operations buyer.

The architecture is the real difference, not the feature list. Lago is a control layer. Billing logic sits close to your product, your data, your team. Chargebee is a managed suite. Billing logic lives inside Chargebee's infrastructure and workflows.

Neither is a value judgment. The managed suite is exactly right for some teams. For others, it's the problem.

The competitive landscape just changed

Before comparing these two: the two biggest pure-play usage-based billing specialists aren't independent anymore.

Stripe completed its acquisition of Metronome in January 2026. Adyen announced the acquisition of Orb on June 11, closed the deal on July 1. If you were evaluating Orb or Metronome as a Chargebee alternative, you're now evaluating a payment processor's roadmap, not a billing product.

Lago stays payment-processor neutral. It integrates with Stripe, Adyen, GoCardless, others, but doesn't route through any of them by default. If your billing strategy includes the right to change processors, that independence is a sharper differentiator than it was 12 months ago.

Deployment: self-hosted vs. cloud-only

The clearest binary between the two platforms. Lago supports managed cloud and full self-hosted deployment. Chargebee is cloud-only, no self-hosted option in public documentation.

For most teams, self-hosting is a preference. For some it's non-negotiable: HIPAA-bound healthcare companies, financial services firms with data residency requirements, government contractors under security review, anyone who needs to certify that billing data never leaves their own infrastructure.

Self-hosted Lago puts you in control of the billing database, the webhook infrastructure, the API integrations, the codebase itself. AGPL means the code is readable and auditable, which matters for financial software where a billing calculation error has real downstream consequences. Migrate off Lago later and your data doesn't move. Your tooling around it does.

With Chargebee, all billing data lives in Chargebee's infrastructure. Not inherently a problem, most SaaS platforms work this way, but worth knowing before you're three years deep in their system.

The migration itself is rarely the hard part. The hard part is data: historical subscriptions, invoices, payment method tokens all need to move or get re-mapped. Teams switching to Lago self-hosted usually run both systems in parallel for a billing cycle, put new customers on Lago first, then backfill existing subscriptions once the rating logic checks out against Chargebee's historical invoices. Budget the parallel-run period. That's the part teams underestimate, not the Lago setup.

Usage-based billing: where the architecture shows

Chargebee handles usage billing fine. This isn't a "Chargebee can't do it" argument. But it was built subscription-first, and usage got layered in later. Lago was built usage-first. Three places that shows up.

Event ingestion. Lago supports REST, batch, SDK, Kafka, Redpanda, Kinesis, and S3 ingestion. Chargebee supports a usage API, bulk import, CSV upload. High event volume or a streaming data pipeline, Lago's options are meaningfully broader. Simple periodic overages, both work fine.

Prepaid credits and wallets. Lago has native wallet infrastructure: credits that expire, roll over, trigger real-time alerts, connect to entitlements. Chargebee's credit model leans toward included-usage allowances and overages, not persistent wallet balances. For AI companies where credit packs are the pricing mechanism, buy 1M tokens, track drawdown, auto-renew at zero, that's a real architectural gap, not a checkbox.

Custom aggregation. Lago supports weighted sum and fully custom aggregation logic. Matters for AI billing where a "token" isn't one unit. Input tokens, output tokens, cached tokens carry different infrastructure costs in pricing tiers like Mistral's. Chargebee covers standard metered patterns but needs workarounds for model-specific token math.

AI billing: the fastest-growing use case

The fastest-growing segment of Lago's customer base is AI and API companies. Same pattern every time: token-based usage with per-model rates, prepaid credit packs, subscription floors for enterprise, custom rate cards negotiated per contract.

Mistral runs its billing on Lago. Per-token rates, separate input and output pricing, volume tiers, pre-purchased token packs for enterprise customers. Exactly the pattern Lago was built to handle.

Three things separate AI billing from standard SaaS billing. Real-time credit enforcement: when a token balance hits zero, API access should change immediately, not at the next invoice cycle. Lago's wallet and entitlement layer does this natively. Chargebee's usage billing is invoice-time, not real-time.

Spend alerts as a product feature: customers using AI products want to see their usage trajectory before they hit the limit. A surprise bill at month-end is a churn risk for an AI company, not just a finance problem.

Frequent repricing: AI compute costs have dropped hard over the last 18 months as models get more efficient. Billing logic locked inside a managed platform's configuration means repricing is a support conversation. Billing logic in Lago's open rating engine means repricing is a config change your team controls.

Payment processor neutrality

Lago has no preferred processor. Stripe, Adyen, GoCardless, others, billing logic stays in Lago regardless of who collects the payment. Route enterprise through Adyen, self-serve through Stripe, no change to your billing configuration.

That matters more now. Metronome is inside Stripe. Orb is now part of Adyen. Buy one of those platforms and you're implicitly buying a payment processor's long-term roadmap. Might be an acceptable trade, both are serious infrastructure, but it's worth being explicit about.

Chargebee supports 30+ processors and isn't processor-captive the way the specialists are becoming. But it's a managed suite: processor changes go through Chargebee's configuration. Self-hosted Lago, you control that layer directly.

Pricing: what you actually pay

Lago's open-source version is free to self-host. You pay for your own infrastructure, Kubernetes, a database, servers, and the engineering time to run it. Lago Cloud is a flat monthly fee that doesn't scale with revenue. $1M ARR or $50M ARR, the bill's the same.

Chargebee's pricing is sales-led, not published. Market expectation is a fee that scales with revenue or transaction volume, exact terms depend on your contract.

The TCO crossover, where Lago Cloud gets cheaper than a revenue-percentage platform, typically lands somewhere in the $3 to 5M ARR range, depending on your Chargebee contract. Below that, managed platforms often win on cost. Above it, flat-fee or self-hosted usually wins.

One honest caveat: self-hosting carries real operational cost. DevOps time, infrastructure provisioning, ongoing maintenance. No existing Kubernetes infrastructure, budget 2 to 4 weeks of setup plus ongoing maintenance before choosing self-hosted purely to save money. For a lot of teams, Lago Cloud is the right place to start.

Feature comparison

When Lago is the right choice

Your billing model is usage-first. AI companies, API platforms, infrastructure tools, anywhere tokens, API calls, compute seconds, or data processed drive revenue. Lago's metering architecture was built for exactly this.

You need self-hosted deployment. Healthcare, financial services, government contractors, or any team with data residency requirements a cloud-only vendor can't satisfy.

You want PSP independence. Multi-processor strategy, changing processors as you expand geographically, or just not wanting billing logic tied to a payment network's acquisition agenda. With the two biggest UBB specialists now inside Stripe and Adyen, Lago is increasingly the only major platform staying neutral.

Billing logic is strategic to you. New AI models, new credit pack structures, new enterprise tiers, you want those changes fast, testable, and in your team's hands, not a vendor's support queue.

You're past roughly $3 to 5M ARR and cost predictability matters. Flat-fee infrastructure gets significantly cheaper than revenue-scaled pricing above that range.

When Chargebee is the right choice

You're subscription-first with usage as a secondary layer. Recurring seats or flat fees as most of your revenue, usage as an overage on top. Exactly what Chargebee was built for, and it does it well.

Finance owns the billing project. Chargebee has native RevRec, ASC 606 and IFRS 15, mature dunning workflows, accounting integrations with NetSuite, QuickBooks, Xero, an admin interface built for operations teams. Finance-led project, Chargebee's buyer profile fits better.

You want one vendor for billing, RevRec, dunning, and payments administration. Chargebee is a broad, genuinely mature suite. One platform covering the full revenue operations surface with SLA-backed enterprise support is a real advantage.

A few common questions

Is Lago a Chargebee replacement, or does it run alongside Chargebee? For most teams, a replacement. Lago handles metering, rating, invoicing, and wallets end to end. Some teams keep Chargebee for a legacy subscription book while new usage-heavy products run on Lago during the transition, but that's a migration state, not a long-term architecture.

How long does switching take? A straightforward subscription model, teams typically go live in 2 to 4 weeks self-hosted, faster on Lago Cloud. Complex custom rate cards or a large historical invoice backlog stretch that. Validating rating logic against your existing invoices is usually the bottleneck, not the setup.

Does Chargebee support self-hosting for large enterprise customers? Chargebee's public documentation shows no self-hosted option. There's a single third-party reference hinting at some enterprise-only exploration, nothing Chargebee has published. Treat it as cloud-only until Chargebee says otherwise.

The honest summary

Chargebee is right for subscription-first SaaS teams that want a mature, finance-owned revenue platform and don't need self-hosting, deep AI credit mechanics, or PSP neutrality.

Lago is right for engineering-led teams building usage-intensive products, AI, APIs, infrastructure, who want billing logic they control, can audit, can deploy anywhere. Mistral and Blacksmith both run on it.

These two platforms aren't really competing for the same buyer. If you've read this far and you're still asking "which one," it comes down to one question: does your billing model need to live inside your engineering stack, or inside a managed SaaS suite?

The former, deploy the open-source version or book a demo to talk through your pricing model. The latter, Chargebee is a good product.