Lago vs Chargebee 2026: Usage-Based Billing Compared

The short answer: pick by who owns billing logic

Lago fits teams where engineering owns billing logic and wants metering, credits and invoices in code and source control. Chargebee fits teams where finance and operations own billing and need dunning, revenue recognition, quoting and a no-code pricing workflow around the usage meter. Most of the decision comes down to that ownership question. Feature checklists matter less.

Usage-based billing is a pricing model where the invoice is calculated from metered consumption events, such as API calls, tokens or seats used, instead of a flat recurring fee. Both products do this. They differ in where they started. Lago started as metering infrastructure, and its own comparison page says it was built for usage from the start. Chargebee started as a subscription management platform and added usage metering on top of that core.

TL;DR: choose Lago when you have engineers to run it, need unusual aggregation logic or want to self-host for data residency. Choose Chargebee when a finance team needs dunning, revenue recognition and quoting to work without writing code. If you are unsure, price both on your real volume before talking to either sales team.

One naming change matters before you compare plans. Chargebee's public pricing now centers on a plan called Flow. Usage Pricing, an independent tracker, reports that Chargebee replaced its Starter and Performance billing tiers with a single Flow plan on August 11, 2026, according to Usage Pricing (2026). Older comparison articles, including some written this year, still describe a free Starter tier, so treat any review that mentions a free entry threshold as out of date. On the Lago side, I found no acquisition or rebrand in the sources I opened. It remains an independent company with a single Premium plan listed on its pricing page.

Key Takeaways

• Lago is open-source (AGPLv3) billing infrastructure with self-hosted and cloud deployment, and its pricing page publishes no dollar figures, so every cost beyond free self-hosting is Quote-based.

• Chargebee now publishes self-serve Flow pricing with two options: $0 plus 0.80% or $99 plus 0.65% of billing volume. Enterprise remains Quote-based.

• On Chargebee's two published options, the break-even sits at about $66K of monthly invoicing. Above that, the $99 plus 0.65% option is cheaper.

• Lago's open-source edition removes the license fee but moves infrastructure, upgrades, monitoring and capacity planning to your team.

• Chargebee's strengths are finance-side: dunning, revenue recognition products, CPQ and 40+ payment gateways. Lago's strengths are engineering-side: flexible metering, ingestion channels and idempotent event APIs.

• Neither product removes the manual reconciliation between metered usage, invoices and cash. That work needs an owner whichever you pick.

Who this is for

This guide is for a finance or engineering lead at a B2B SaaS company that already knows it needs usage-based or hybrid pricing and is close to selecting a billing system. It assumes you have a pricing model in mind, such as per-token, per-API-call, committed-spend or seat-plus-overage.

Red flags that should rule one option out:

• Red flags for Lago: no engineer can own a service in production, you need native revenue recognition at close time this quarter, or your auditors require a vendor-supported dunning and collections workflow you cannot build.

• Red flags for Chargebee: your pricing needs custom aggregation that standard metering patterns cannot express, you require self-hosting for data residency, or you cannot accept a percentage-of-billing fee as volume grows.

How we evaluated Lago and Chargebee

This is an informed buyer's guide built from public information. We did not run either product, so every capability below comes from vendor documentation, vendor pages or independent write-ups, and vendor-authored claims are labelled as such. The criteria below are the ones a usage-pricing buyer typically decides on. The weights are our analysis and reflect a reader who is mid-evaluation, not a universal ranking.

Evidence quality varies. Chargebee's Gartner placement comes from Chargebee's own materials. Lago's comparison pages are written by Lago, and the competitor commentary we cite from Orb is written by a direct Lago rival. We say which is which wherever it matters.

Side-by-side feature matrix

The matrix below normalizes the same questions across both products. Cells reflect the vendors' own pricing pages and documentation unless marked as a vendor claim.

Sources for the matrix: Lago's pricing page, Chargebee's pricing page, Lago's own comparison and Orb's Lago review. Because Chargebee's page lists its pending-state invoices, usage thresholds and overage handling as included in both Flow and Enterprise, the finance review step for metered invoices is a native feature there, while on Lago it is the draft invoice state and grace period management that serve the same purpose.

What each costs, with the checked date

Pricing checked October 12, 2026. The only dollar figures on either vendor's page are Chargebee's Flow figures. Lago publishes none.

The headline number for Chargebee is the Flow structure. Flow second option: $99/month plus 0.65% of billing, 100M events included according to Chargebee (2026). The pricing page shows the two formulas next to a volume slider but does not label which belongs to which option. Rework's 2026 Chargebee comparison names them Pay As You Go and Commit Monthly and confirms the same rates, which is why we map the $0 formula to the pay-as-you-go option. Confirm that mapping in the checkout flow before you rely on it.

The two Flow options cross where the monthly fee is paid back by the lower rate. The difference is 0.15 percentage points, so $99 divided by 0.0015 gives $66,000 of monthly volume, which matches the slider's $66K breakeven label.

Those totals exclude paid add-ons, taxes and any negotiated Enterprise discount. Lago's page, according to Lago (2026), lists Premium as Contact us with add-ons on demand, including 24x7 incident management and an assigned solution engineer, so a like-for-like Lago cost is Quote-based. If a vendor or a review gives you a Lago dollar figure, ask for the order form.

The structural point matters more than any single number. Chargebee's cost scales with your invoiced revenue. Lago's open-source edition scales with your infrastructure and your engineers, and its paid edition is priced by volume and quoted per customer according to Lago's own comparison page. A fast-growing usage-based business should model both curves at two or three future volumes, not at today's.

A worked example: $200K a month and 120M events

Take an illustrative company that invoices $200,000 a month and sends 120M usage events a month. On the Chargebee $99 plus 0.65% option the monthly cost is $99 + (0.65% × $200,000 = $1,300) = $1,399, or $16,788 a year, but 120M is 20M above the 100M events the plan includes, so the overage add-on is Quote-based and sits on top of that figure. For Lago's open-source edition, assume $1,000 a month of infrastructure ($12,000 a year) and a quarter of one engineer at an assumed $180,000 loaded cost ($45,000 a year), which totals $57,000 a year before any Premium add-on quote. Those assumptions are placeholders to replace with your own numbers. The integration work also differs: on Lago each event posts to POST /events with a transaction_id that doubles as the idempotency key, while on Chargebee a usage record carries item_price_id and usage_date and only one record is allowed per subscription, item price and date, so a daily roll-up is common before sending. The invoice signals differ in name as well, with Lago emitting invoice.drafted for a draft and Chargebee emitting invoice_generated for a finished invoice.

The table does not say Chargebee is cheaper in general. Self-hosting amortizes across more engineering value if your pipeline already runs on Kafka or S3 and the engineers exist anyway. The Chargebee side also omits what its Quote-based RevRec and event add-ons would cost. The point is that the cheapest-looking line on a pricing page is rarely the cheapest total.

Lago: best fit, limits and implementation

Best fit: engineering-led companies whose pricing logic is unusual or changes often. Examples are AI products billing input, output and cached tokens at different rates, or marketplaces taking a percentage with fixed fees and caps. Lago's pricing page lists billable metrics, real-time usage alerts, prepaid credits and wallets, plan overrides per customer, progressive billing, credit notes and an invoice preview endpoint. Its ingestion is built around an events endpoint. According to Lago (2026), each event needs three required fields (transaction_id, external_subscription_id and code), and transaction_id is the idempotency key. That design is what lets an engineer replay a failed batch without double-charging. The same docs warn that on the ClickHouse-based pipeline, an event with the same transaction_id and timestamp overwrites the existing one instead of being rejected.

Vendor evidence is reasonable but not extensive. Lago total funding: $22M by March 2024 according to Pulse 2 (2024), with a $15 million Series A led by FirstMark and a seed round before it. I found no later round in the sources I opened.

Limitations come from three places. First, finance depth: Lago's own comparison concedes that its dunning is supported but less polished than Chargebee's, and Orb describes revenue recognition and quotes as paid add-ons, with revenue recognition labelled Alpha. Second, packaging: some capabilities sit behind paid add-ons, so the open-source edition alone will not give finance a complete tool. Chargebee's own comparison page claims the free self-hosted tier excludes the customer portal, credit notes, dunning and tax integrations. That is a competitor's claim, so verify it against Lago's current docs. Third, operations: Orb notes that self-hosting places infrastructure, upgrades, monitoring, security and capacity planning on the deploying team, according to Orb (2026). That claim comes from a Lago competitor, but it is also simply what self-hosting means.

Implementation: expect an engineering project. You define billable metrics, wire your application to emit events, connect a payment provider and an accounting system, and decide how drafts become finalized invoices. Lago's webhook reference uses invoice.drafted for a draft and invoice.created for a finalized invoice, which is easy to misread, so test your event handlers against the reference rather than assuming the names mean what Stripe's do.

Choose Lago if you have engineers who will own a billing service, need data to stay in your own infrastructure, or have aggregation logic that standard metering cannot express. Look elsewhere if you need finance-owned tooling out of the box.

Chargebee: best fit, limits and implementation

Best fit: finance-owned billing where usage is one part of a broader subscription motion. Chargebee's pricing page lists usage ingestion through API, CSV or S3, custom SQL metering, tiered, stair-step, volume and block pricing, hybrid recurring-plus-usage invoices, and pending-state invoices that can be reviewed manually or auto-closed. It also lists real-time usage limits and alerts, and connections to 40+ payment gateways. Around billing it sells CPQ, RevRec and Growth products, each with its own Quote-based tiers.

On analyst evidence, Chargebee says in a vendor-authored post that it was named a Leader in Gartner's 2026 Magic Quadrant for Recurring Billing Applications, its third consecutive year, according to Chargebee (2026). The report is dated August 12, 2026. We have not seen the report itself, so treat the placement as a vendor claim until you obtain a copy.

Limitations: the cost model and the fit with exotic usage logic. A percentage of billing volume grows with your revenue, and the 100M monthly events included in Flow are a hard planning number for high-volume event streams, because the 500M tier is a paid add-on. Lago's comparison page claims Chargebee needs workarounds for model-specific aggregation, such as token pricing with different rates per token type. That is a competitor's claim, so test your own pricing model in Chargebee's sandbox before assuming either side. Chargebee's usages API reference documents specific constraints, including one usage record per date for each subscription and item price, and says legacy metered billing caps usages per subscription over its lifetime. For high-volume ingestion it points to the separate Usage Events API. See the usages API reference for the current rules.

Implementation: Rework's 2026 comparison reports that finance-led teams with clean data may go live in 30 to 45 days, while migrations involving usage billing or messy legacy data often take longer. Usage-billing migrations: often 10 to 12 weeks according to Rework (2026). Rework notes those timelines come from general guidance, not an audited study. The heavier work, it says, is data migration plus tax and gateway configuration, not the core setup. If you are weighing Chargebee against other subscription platforms, our guides on Recurly versus Chargebee and Maxio for Series A SaaS and Chargebee versus Recurly cover the subscription-first comparison. If you already run on Stripe, Stripe Billing versus Chargebee is the closer read.

Choose Chargebee if finance owns the process, you need dunning, revenue recognition and quoting from one vendor, and your pricing fits standard metered patterns. Look elsewhere if you need self-hosting or your fee exposure at higher volumes is a deal-breaker.

Where the manual work still lives

Neither product closes the loop between what your application metered, what the invoice says and what finance approves. That gap is usually filled with spreadsheets, scripts and Slack messages. The following is a proposed, configurable workflow, not a description of a live deployment or a measured result. With US Tech Automations, the trigger would be a draft or pending invoice signal: invoice.drafted from Lago or invoice_generated from Chargebee. The action would pull the invoice's usage totals through the vendor API or a data export, compare them with the event counts from your own product database, and flag any customer whose variance exceeds a threshold finance sets. The output would be a reconciliation row per flagged customer and a review task for a named finance owner, who approves, corrects or holds the invoice. Prerequisites are API credentials for the billing system, read access to your event store or a warehouse export, and a written threshold. The human review point is the approval itself, since the workflow would not finalize or void anything on its own.

A second proposed workflow covers collections. The trigger would be an overdue or failed payment signal, such as invoice.payment_overdue in Lago or payment_failed in Chargebee. The action would look up the account's contract tier and last contact in your CRM, then draft a tiered reminder or an internal escalation to the account owner. The output is a logged message and a task. A person approves any message to a strategic account before it sends, and the design would need CRM access plus a rule for which accounts are exempt from automated contact. Both flows write each step to a run log, which gives finance an audit trail separate from the billing system.

The fair alternative is stitching this together yourself in Zapier, Make or n8n, or building it in-house. Those tools can support run histories, retries, error branches and audit evidence when they are configured well. The cost is that you design and own the observability, idempotency on replays, escalation paths, access controls and ongoing maintenance of every flow. A proposed US Tech Automations design would configure the same pieces as one reviewed workflow, with idempotency keys on the reconciliation records, a named escalation owner per exception type and role-scoped credentials, but it still requires the API and export access above and still needs your team to set thresholds and review exceptions.

Decision checklist

Answer these before you request a quote from either vendor:

• Who will own billing logic day to day: engineering, finance, or a shared RevOps team?

• Can your pricing model be expressed with standard metered patterns, or does it need custom aggregation?

• What are your monthly invoiced volume and event count today, and in 12 and 24 months?

• Do you need self-hosting or strict data residency?

• Does your auditor expect native revenue recognition, and by when?

• Which payment gateways and ERP must you connect, and are they supported by name?

• Do you have a named engineer for a self-hosted service, including on-call?

• What is your plan for migrating historical subscriptions and usage?

Common mistakes

• Treating free software as a free system. The open-source edition removes a license line but adds infrastructure, upgrades and engineering.

• Comparing a percentage fee to a flat fee at today's volume only. The Flow break-even of about $66K a month shows how fast the answer can flip.

• Trusting an old review. Several 2026 articles still describe a free Chargebee tier that Usage Pricing reports was replaced in August 2026.

• Assuming webhook names behave like another vendor's. Lago's invoice.created means finalized, not drafted.

• Skipping the event ceiling. Chargebee's included 100M monthly events is a hard planning figure, and Lago's docs state no rate limit, so test both at your peak.

• Reading vendor comparison pages as neutral. Both vendors publish pages that favor themselves, and Orb is a competitor to Lago.

Glossary

• Billable metric: the rule that turns raw events into a quantity to bill, such as a sum or unique count.

• Idempotency key: an identifier, such as Lago's transaction_id, that lets a retried event be recognized instead of billed twice.

• Pending invoice: an invoice held in a review state before finalization, so finance can inspect usage first.

• Dunning: the sequence of retries and reminders that recovers failed or overdue payments.

• Revenue recognition: accounting treatment that decides when billed amounts count as revenue, under standards such as ASC 606.

• CPQ: configure, price, quote software for building and approving customer quotes.

• Prepaid credits: a balance the customer buys up front and draws down as usage accrues.

• AGPLv3: the open-source license on Lago's core, which carries obligations if you modify and offer it as a network service.

Frequently asked questions

Is Lago cheaper than Chargebee?

It depends on volume and engineering time, because Lago's open-source edition has no license fee but costs infrastructure and people, and its paid editions are Quote-based. Chargebee's published Flow options cost $0 plus 0.80% or $99 plus 0.65% of billing volume, so you can compute its price today and cannot compute Lago's without a quote.

Is Lago really open source?

Yes, its core is released under AGPLv3 with code on GitHub, according to Orb's review, though several finance features are sold as paid add-ons. Check which features your plan includes before you assume the free edition covers them.

Does Chargebee support self-hosting?

Not according to Lago's comparison page, which says Chargebee is cloud-only with no self-hosting option in its public documentation. That is a competitor's statement, so ask Chargebee directly if residency is a requirement.

Which handles AI token pricing better?

Lago markets weighted-sum and custom aggregation for token pricing, while Chargebee lists custom SQL metering and tiered, stair-step and volume models. Both claims are vendor-authored, so model your actual token types, such as input, output and cached tokens, in each sandbox before deciding.

Do either of them do revenue recognition?

Chargebee sells a separate RevRec product with Quote-based tiers, and Lago has revenue recognition as an add-on that Orb describes as Alpha. If native ASC 606 handling is a hard requirement this year, that difference should weigh heavily.

When NOT to use US Tech Automations

A simpler tool wins in a few cases. If your billing runs through a single system with few exceptions, the vendor's own pending-invoice review and dunning settings are enough, and an added workflow layer only adds something to maintain. If nobody can provide API credentials or a data export from the billing system, a reconciliation workflow has nothing to read. And if your team already runs a well-maintained n8n or Make setup with run logging and a named owner, rebuilding it elsewhere is rarely worth the migration.

Where to land

If your engineers want billing as code and your finance needs are modest, shortlist Lago and ask for a Premium quote alongside a self-hosting estimate. If finance owns the process and your usage patterns are standard, shortlist Chargebee and model the Flow options at your real volume, including the event ceiling. In either case, decide who owns the gap between metered usage, invoice approval and collections before you sign. If you want to see how that review-and-reconcile layer could sit on top of whichever you choose, see how US Tech Automations configures this.