Usage-based billing for AI database agents: why we chose Polar.sh

As we move EnginiQ from a core runtime toward a full-scale control plane, one of the most important components we had to design was our monetization model.

Building a developer tool for AI agents presents a unique challenge for billing: how do you charge fairly for a system that can run thousands of operations in seconds, but where safety and trust are the primary value drivers?

The "Safe Execution" Metric

In most SaaS products, you charge per seat. But for EnginiQ, the value isn't just in how many people use the dashboard—it's in how many safe, audited operations the agents perform.

We decided on tool execution counts as our primary usage metric.

• Hobby Users: Get a generous free tier of safe schema discovery and local dev tools.
• Pro Teams: Benefit from higher limits, dedicated audit storage, and team-based approval workflows.

Why Polar.sh?

When it came to implementing this, we chose Polar.sh as our payment and subscription partner. For a developer-first product, Polar offers three critical advantages:

1. Developer-First SDK

Integration was seamless. Using the @polar-sh/sdk, we were able to implement subscription checks directly into our core engine's execution lifecycle in under an hour.

2. Built-in usage grants

Polar's "Benefit" system allows us to link specific products (like our Pro plan) directly to access rights in our system. This means when a user upgrades on Polar, their usage_limit is automatically updated in our database via their secure webhook system.

3. Modern, lean interface

We want EnginiQ to feel modern and fast. Polar's checkout and customer portal align perfectly with our aesthetic of "premium, dark-mode, high-performance tooling."

How it works under the hood

In the latest v0.1 update, every time an agent calls runTool, the EnginiQ core performs a three-step check:

1. Identity: It identifies the actor_id (the developer or specific agent).
2. Subscription: It checks our local enginiq_subscriptions cache (synced with Polar).
3. Usage Enforcement: It counts successful tool runs for the current billing period using our AuditLogger.

If a user hits their limit, the engine doesn't just crash—it returns a structured USAGE_LIMIT_REACHED error with a link to their Polar checkout page to upgrade.

Building for the long term

Monetization isn't just about revenue; it's about building a sustainable ecosystem. By choosing a usage-based model powered by Polar, we ensure that as agents get more powerful and run more operations, EnginiQ can continue to scale its security and audit infrastructure alongside them.

We're excited to see how teams use the Pro tier to secure their production Postgres workflows.

---

[!TIP] You can preview the billing management interface right now in our Hosted Dashboard.