Lago

Usage-based and subscription billing, honestly reviewed. No marketing fluff, just what you get when you self-host your revenue stack.

TL;DR

• What it is: Open-source (AGPL-3.0) metering and usage-based billing platform — think Chargebee or Stripe Billing, but the code runs on your infrastructure and the vendor has no visibility into your billing data [README][3].
• Who it’s for: Engineering-led startups and scale-ups with complex pricing models — usage-based, hybrid, prepaid credits, custom contracts. Not a “set up in 10 minutes” tool for non-technical founders [4].
• Cost savings: Chargebee and Recurly charge a percentage of revenue plus monthly fees that compound as you grow. Lago self-hosted is free. The math is obvious once your MRR clears $10K [3][4].
• Key strength: The most complete open-source billing engine available. Event-driven architecture that handles usage metering, subscriptions, invoicing, prepaid credits, and coupons in one system. Trusted by Mistral AI, PayPal, Groq, and 1NCE [README][homepage].
• Key weakness: AGPL-3.0 license (not MIT) means embedding it in a commercial product requires careful legal review. The community edition is genuinely capable, but pricing for the cloud or enterprise tier is opaque — “contact us” everywhere, which one AlternativeTo reviewer called “the worst possible pricing model” for an open-source product [1].

---

What is Lago

Lago is an API-first billing infrastructure layer. You send it events — API calls, compute minutes, messages processed, seats used — it applies your pricing rules, manages subscriptions, and generates invoices. The founders describe it as “the open-source alternative to Stripe Billing and Chargebee,” which is both accurate and undersells the complexity of what it actually handles [3][README].

The company was founded by people who ran revenue at fintech companies before. The pitch in their Product Hunt launch [3] is telling: “After owning revenue at a Fintech unicorn before (from pre-launch to $100M ARR), we learnt first hand that pricing optimizations were very much slowed down (when not stopped) by clunky billing systems.” They built the thing they couldn’t find in the market — a billing system that’s open, programmable, and doesn’t take a revenue cut.

The project is backed by Y Combinator and sits at 9,447 GitHub stars with the AGPL-3.0 license. It’s written in Ruby on the backend with a React frontend, and the architecture is built around ClickHouse for event ingestion at volume [AlternativeTo tags][5]. As of this review, customers using it in production include Mistral AI, PayPal, Groq, Synthesia, 1NCE (IoT billing), and Laravel [homepage].

What separates it from simpler invoice tools is the event model. You track anything billable as an event, attach pricing rules to those events, and Lago handles the aggregation, proration, and invoice generation downstream. This is the same pattern that Segment, Algolia, and Stripe’s own metering infrastructure use internally [README].

---

Why people choose it

The core argument isn’t about features — it’s about control and cost trajectory.

Versus Stripe Billing. Stripe is the default billing choice for developers because it handles payments and billing in one place. But Stripe’s cut (0.5–0.8% of revenue processed through Billing) becomes a meaningful line item once you’re past $100K MRR. More importantly, Stripe’s pricing model is optimized for Stripe’s interests — it gently pushes you toward Stripe as your payment processor, and complex usage-based pricing requires their Meters API that wasn’t designed for arbitrary event schemas. Lago explicitly positions itself as “payment-agnostic” — it generates invoices and manages subscriptions, but you bring your own payment processor (Stripe, Adyen, GoCardless, whatever) [README][3].

Versus Chargebee and Recurly. These are the SaaS billing incumbents. They’re polished, well-supported, and expensive at scale. Chargebee’s launch plan starts around $249/mo, and the fees scale with revenue. Recurly is similar. Both are closed-source black boxes — you can’t inspect how proration works, you can’t fork the invoicing logic for edge cases, and if they decide to sunset a feature or raise prices, you’re stuck [3][4]. The Alguna comparison blog [4] notes that Lago’s primary advantage here is “maximum control at minimal cost” for teams willing to own the ops.

Versus building in-house. The Lago blog [2] links to a Hacker News thread with 350+ comments from engineers at Qonto, Algolia, Pleo, and Segment describing billing infrastructure nightmares. The recurring theme: billing systems look simple (subscriptions + invoices) until they aren’t. Proration edge cases, dunning logic, tax handling, credit notes, prepaid wallets, multi-currency — every piece is individually manageable and collectively a 12-month engineering project. Lago is the case that you shouldn’t build this yourself.

The AGPL caveat. One AlternativeTo reviewer flagged this directly: “The worst possible pricing model ‘contact us’ while open-source version is barebones (and AGPL, wtf)” [1]. That’s a sharp critique worth unpacking. AGPL-3.0 means if you modify Lago and expose it over a network (which you will — it’s a billing API), you’re required to make your modifications open source. That’s fine for most self-hosters who aren’t building a billing SaaS on top of Lago. It’s a real legal concern if you’re building a multi-tenant product where Lago is a core infrastructure component. The reviewer’s “barebones” characterization of the community edition is contested by the feature list — the self-hosted version is fully functional — but the “contact us” pricing wall for cloud and enterprise tiers is a genuine friction point for teams that want cost predictability before committing.

---

Features

Based on the README, website, and third-party descriptions:

Usage metering:

• Event-based ingestion: send any billable event via API, Lago handles deduplication and aggregation [README]
• Multiple aggregation methods: sum, count, unique count, max, weighted sum [README]
• Real-time and batch processing — one customer aggregates in real-time, another does it monthly, same infrastructure [2]
• ClickHouse backend for high-volume event storage [5]

Pricing and plans:

• Pay-as-you-go, subscription, and hybrid plans [README]
• Custom contracts for enterprise deals [homepage]
• Prepaid credits — charge upfront, draw down as usage accumulates [README]
• Entitlements management [homepage]
• Coupons and promotional credits [README]
• Add-ons: one-time charges invoiced on the fly without waiting for the billing cycle [README]

Billing and invoicing:

• Automatic invoice generation based on plan configuration [README]
• Credit notes and adjustments
• Dunning — automated payment retry and failed payment workflows [homepage]
• Alerting — usage threshold notifications before customers hit limits [homepage]
• Multi-currency support [homepage]

Revenue operations:

• Revenue analytics [README]
• One source of truth for billing data across engineering, product, finance, and ops teams [homepage]

Integrations:

• Payment-agnostic — works with Stripe, Adyen, GoCardless, and others [README]
• REST API for all operations [merged profile]
• Connects to accounting systems, CRMs, ERPs, tax tools [homepage]

Enterprise / self-host:

• Docker and Docker Compose deployment [merged profile]
• RBAC — role-based access control [homepage]
• SOC 2 Type II on the cloud version [homepage]
• Flexible deployment: on-prem, VPC, or cloud [homepage]
• Lago Embedded — white-label billing for platforms [homepage blog link]

---

Pricing: SaaS vs self-hosted math

What Lago charges:

The self-hosted community edition is free. AGPL-3.0 license, no runtime fees, no revenue percentage [README].

For the cloud version and enterprise self-hosted support, Lago uses “contact us” pricing across the board. There are no published tiers on the pricing page. This is the legitimate frustration behind the AlternativeTo comment [1] — for a team doing early evaluation, not knowing whether the cloud version costs $200/mo or $2,000/mo before a sales call is a real barrier.

What you’re replacing:

• Stripe Billing: 0.5% of revenue billed through Billing, plus Stripe’s payment processing fees (2.9% + $0.30 per transaction). At $50K MRR billed through Stripe, that’s $250/mo just for the billing layer, before processing.
• Chargebee Rise plan: $249/mo for up to $100K MRR, then scales to $549/mo, then revenue-based pricing.
• Recurly: similar tiered structure, percentage-of-revenue components at higher volumes.

Self-hosted math:

You need a VPS that can handle your event volume. A $20–40/mo Hetzner server handles moderate traffic comfortably. The Lago stack (Rails backend, React frontend, PostgreSQL, ClickHouse, Redis) has more moving parts than a simple web app — budget for a more capable server than the $6 Hetzner minimum.

At $50K MRR, replacing Stripe Billing’s 0.5% layer with Lago self-hosted saves $250/mo — $3,000/year — and that scales linearly with revenue. At $500K MRR, you’re saving $2,500/mo. The VPS cost is rounding error at that point [4][3].

The catch: this math only works if you have an engineer who can own the billing infrastructure. Lago isn’t a “set it and forget it” tool — billing migrations, schema updates, and the self-hosted operational model require someone who knows what they’re doing [2].

---

Deployment reality check

Lago is honest about self-hosting complexity in a way that most open-source tools aren’t. Their engineering blog [2] describes the internal challenge directly: “Building a self-hosted product is harder… We have zero visibility [into customer instances]. We can’t know when a migration is complete.” Something as simple as adding a database column can break a billing run mid-cycle for self-hosted customers.

What you actually need:

• Linux server with sufficient RAM for ClickHouse event storage (4GB minimum, 8GB+ recommended for production)
• Docker and Docker Compose
• PostgreSQL (bundled or external)
• Redis (bundled or external)
• ClickHouse for event aggregation
• A reverse proxy (Caddy or nginx) with HTTPS
• An SMTP provider for invoicing and dunning emails

What the stack looks like: Lago is a multi-container application. The Docker Compose file brings up the Rails API, the React frontend, Postgres, Redis, ClickHouse, and a background worker. That’s six services. It’s not unusual for billing infrastructure, but it’s notably heavier than, say, deploying Plausible Analytics.

What can go sideways:

• Database migrations during live billing runs are the documented failure mode [2]. Lago’s engineering team ships updates that require careful migration sequencing — self-hosters who auto-update without reading the changelog can break invoice generation.
• ClickHouse is the least common component in a typical developer’s stack. If it fails or fills disk, event ingestion stops silently. You need monitoring.
• The AGPL license means you need to evaluate your use case before putting it in a commercial product. Not a deployment problem, but a business problem that surfaces at deployment time.
• The AlternativeTo reviewer’s characterization of the open-source version as “barebones” doesn’t match the feature list, but the cloud/enterprise gap (SLA, support, managed migrations) is real. For a startup using it for its own billing, the community edition is complete. For a company whose billing system downtime means revenue loss, the lack of a paid support tier with a published price point is a genuine risk.

Realistic time estimate for a developer familiar with Docker: 2–4 hours to a working instance. For production readiness including monitoring, backup configuration, and SMTP: 1–2 days. For a non-technical founder with no Linux experience: this isn’t the tool to start with — Lago Cloud or a deployment service is the right path.

---

Pros and Cons

Pros

• Event-driven architecture designed for complex pricing. The metering model handles anything you can track as an event. Usage-based, subscription, hybrid, prepaid — all first-class, not afterthoughts [README][3].
• Payment-agnostic. Lago generates invoices and manages subscriptions; you keep your existing payment processor. No vendor lock-in to a specific payment rail [README].
• No revenue percentage. Chargebee and Recurly both have revenue-based pricing that scales against you. Lago’s self-hosted version is flat cost [3][4].
• Trusted at scale. Mistral AI, PayPal, Groq are production customers. This isn’t an experimental tool — it handles serious billing volumes [homepage].
• Full data ownership. Every invoice, every metered event stays in your infrastructure. For fintech, healthcare, or any regulated industry, this is the feature [2][4].
• REST API for everything. The full billing surface is programmable. Automation, custom dashboards, RevOps tooling — it’s all possible [merged profile].
• Actively developed by a YC-backed team with real customers on it. Not a one-person project at risk of abandonment [3].

Cons

• AGPL-3.0, not MIT. For a company building a multi-tenant platform on top of Lago, the copyleft requirement is a real legal constraint. Get legal review before embedding this commercially [1][README].
• “Contact us” cloud pricing. No published tiers for the managed version. One reviewer called it out explicitly [1]. Evaluation friction for teams that want cost predictability.
• More infrastructure than you might expect. Six Docker services including ClickHouse is non-trivial to operate, monitor, and upgrade safely [2].
• Self-host operational complexity is documented and real. The Lago team wrote a blog post about how hard it is to ship updates to self-hosted customers without breaking things [2]. That honesty is admirable; it’s also a warning sign that the operational burden is higher than average.
• Not for non-technical teams. The UI is capable, but Lago is API-first by design. If your finance team wants to own billing iterations without engineering involvement, the learning curve is steep [4].
• Community edition support is community-only. Slack and GitHub issues. No SLA. For a billing system that processes revenue, “ping us on Slack” support is a business risk.
• Limited third-party reviews. Unlike established tools with hundreds of G2 or Capterra reviews, Lago’s review surface is thin — Product Hunt, AlternativeTo, and a handful of comparison blogs. The signal-to-noise ratio in those sources is lower [1][3].

---

Who should use this / who shouldn’t

Use Lago if:

• You’re a developer-led startup with usage-based or hybrid pricing, and the billing complexity is real enough that Stripe Billing’s model feels like a constraint.
• You’re past the point where billing system costs (percentage of revenue) are a meaningful line item and you want to own the infrastructure.
• Your industry requires data residency or on-premise deployment — fintech, healthcare, government procurement.
• You want to embed custom billing logic that closed-source vendors won’t support.
• You have the engineering capacity to operate a six-service Docker stack and treat billing infrastructure as a first-class system.

Skip it if you’re non-technical:

• If you’ve never set up a multi-service Docker application, this is not the place to start. Use Lago Cloud (with its “contact us” caveat acknowledged) or find a deployment partner.

Skip it (use Stripe Billing) if:

• Your billing model is simple — flat subscriptions or one-time charges. Stripe Billing handles that cleanly and the integration complexity is near-zero.
• You’re pre-revenue or under $5K MRR. The setup cost doesn’t pay off at that scale.

Skip it (use Chargebee or Recurly) if:

• You need a non-technical-friendly UI for a finance team that runs billing without engineering support.
• You want enterprise SLAs with published pricing and phone support.
• The AGPL license is a dealbreaker for your use case.

Skip it (evaluate Kill Bill) if:

• You need Apache-licensed billing infrastructure with a longer track record. Kill Bill is older, more battle-tested in enterprise contexts, and Apache 2.0 [5].

---

Alternatives worth considering

From the data and comparison articles:

• Stripe Billing — the incumbent default. Easiest integration if you’re already on Stripe, 0.5% fee on billed revenue, closed-source, pushes you toward Stripe as your payment processor.
• Chargebee — polished SaaS billing, strong finance team UX, handles complex subscription logic, no self-host option, $249+/mo and scales with revenue.
• Recurly — similar positioning to Chargebee, longer track record in subscription billing, enterprise-focused, closed source.
• Kill Bill — Apache 2.0 licensed, Java-based, open-source billing with a long enterprise track record, steeper learning curve and older UI [5].
• Alguna — newer quote-to-revenue platform with CPQ, revenue recognition, and non-technical UI. The comparison blog [4] frames it as the “next step after Lago” for teams that need GTM motions alongside billing.
• Meteroid — open-source, listed as an alternative on AlternativeTo [1]. Newer and less battle-tested than Lago.
• In-house build — the option that the 350+ Hacker News comments in Lago’s launch thread [3] collectively argued against. The failure mode is well-documented.

For a company with genuine usage-based pricing complexity and the engineering capacity to operate it, the realistic shortlist is Lago vs Stripe Billing vs Chargebee. Choose Lago if data sovereignty, AGPL permissiveness for your use case, and no revenue-based fees matter. Choose Stripe Billing if simplicity of integration outweighs cost. Choose Chargebee if your finance team needs to own billing without engineering in the loop.

---

Bottom line

Lago is the most credible open-source answer to the billing infrastructure problem that engineering teams have been solving badly for 20 years. The event-driven metering architecture is genuinely well-designed, the feature set covers the full billing lifecycle, and the customer list — Mistral AI, PayPal, Groq — validates that it handles real production load. The trade-offs are honest: AGPL-3.0 creates licensing considerations for commercial embedding, the self-hosted stack has real operational complexity, and the “contact us” cloud pricing is a friction point for teams doing early evaluation. But for an engineering-led startup with hybrid pricing and a team that can own infrastructure, the calculus is clear — Stripe’s 0.5% billing fee and Chargebee’s revenue-scaled pricing both compound against you as you grow. Lago’s self-hosted version compounds at $0.

If the operational complexity is the blocker, that’s exactly the kind of deployment work that upready.dev handles for clients.

---

Sources

1. KazimirPodolski, AlternativeTo — Lago Review (1 review, Dec 2025). https://alternativeto.net/software/lago/about/
2. Finn Lobsien, Lago Blog — “Why building a self-hosted SaaS is a headache (and how we make it easier)” (Aug 5, 2025). https://www.getlago.com/blog/self-hosted-saas
3. Raffi Sarkissian et al., Product Hunt — “Lago — Open-source alternative to Stripe Billing and Chargebee” (launch post, 4.6/5, 10 reviews). https://www.producthunt.com/products/lago
4. Alguna Blog — “Lago alternatives: Why modern SaaS teams switch to Alguna”. https://blog.alguna.com/lago-alternatives/
5. Awesome Open Source — “Lago Alternatives” (project metadata, Stars, License, Categories). https://awesomeopensource.com/project/getlago/lago

Primary sources:

• GitHub repository and README: https://github.com/getlago/lago (9,447 stars, AGPL-3.0 license, YC-backed)
• Official website: https://www.getlago.com
• Hacker News launch thread: https://news.ycombinator.com/item?id=31424450 (350+ comments)