Challenge

Chipp.ai launched in 2023 with the goal of democratising AI access. Based in Fargo, North Dakota, the startup wanted to help smaller businesses create the kind of custom, brand-native AI agents that large corporations were building for their websites. So, it built a no-code platform that allows customers to develop, train and operate their own AI agents. With Chipp, businesses like hotels, medical practices and agencies can mix and match popular LLM models from Anthropic, OpenAI and Google – without paying for multiple subscriptions.

At first, Chipp billed customers on a traditional SaaS subscription model. But this setup quickly became a mismatch: Chipp's expenses could swing depending on how many LLM tokens each customer consumed – both while building the interface and after agents went live. A small number of power users could generate outsized costs that ate into Chipp's profit margins, like one business that ran up a $4,000 token bill on a $29 monthly plan. On top of that, Chipp's team had to spend hours manually creating invoices and following up on payments.

"We basically were trying to guess what subscription to charge to make sure we didn't lose money," said Scott Meyer, Chipp cofounder and CEO. "It became clear that we didn't want a business model that's antithetical to our customer's success."

Chipp considered building an in-house token tracking and billing system, but that would have pulled the engineering team away from time spent on improving the product – and it would have required ongoing updates as models and token pricing changed.

Instead, Chipp needed a usage-based billing solution that it could implement quickly – one that provided clear visibility into customer LLM usage, protected margins and helped customers make better decisions about which models to use.

Solution

Chipp was already using Stripe, having integrated Stripe Billing, Stripe Payments and Stripe Tax to launch globally from day one. The company had also implemented Stripe Connect, so agencies and consultants could build AI agents for their customers and pass through the cost of using the Chipp platform. In December 2025, Chipp updated its Stripe integration to become one of the first to use Stripe's usage-based billing feature for LLM tokens.

Setting up the new system was simple: because Stripe documentation is formatted for LLMs, Chipp cofounder and chief product and technology officer Hunter Hodnett used Claude Code to configure the company's Billing account for LLM token billing. This usage-based layer captures LLM requests, routes them through Stripe's AI gateway to record tokens consumed for each model, aggregates usage, and ties it back to the customer's unique Stripe ID. Stripe also tracks changes in token costs and updates the Chipp billing logic accordingly.

With token tracking capabilities in place, Chipp updated its subscription pricing with a simple, transparent formula: businesses on Chipp receive one-third of their monthly subscription fee in tokens and then pay for additional tokens used beyond that limit – with Chipp charging a 30% markup to cover its business costs. Agency businesses can white label Chipp and offer their customers the same pricing model, and collect payments for token usage through Connect.

To avoid customers unknowingly racking up bills, Hodnett used the Stripe billing alerts API to set up a webhook to automatically notify customers when their tokens run low and prompt them to pay for additional tokens. If customers reach their limit, the system freezes their LLM access until they purchase more.

Chipp also adopted Stripe Sigma to augment the native reporting in usage-based billing and help to analyse multiple areas of its business. Stripe Sigma allows businesses to analyse data by building custom reports, SQL prompts and natural language queries. Chipp uses Stripe Sigma to gain deeper insights about the impact of pricing changes and ancillary offerings such as paid courses.

Results

Two days to implement token-based billing

With Stripe's LLM-friendly documentation and collaborative support, Chipp implemented token-based billing in just two days.

"Token-based billing saved our business. Building this ourselves would have taken us a few months that we didn't have," said Hodnett.

Usage-based billing also eliminated the need to manually update pricing for changes in models and token pricing, which would have been a significant time investment.

Token-based billing generates 20% revenue lift

Switching to a usage-based model aligned Chipp's business model with its customers' success, generating more revenue the more its customers used the product. Chipp saw a 20% increase in revenue in the first month after adopting the new billing model, and over 5.9 billion tokens billed within the first 2 months.

Automated billing eliminates time spent on manual invoicing

Thanks to the low-token alert system created with the billing alerts API, Chipp eliminated manual billing and invoicing processes that had been eating up roughly 25% of the team's time. Instead of cumbersome manual workflows, Chipp now automatically tracks token usage through Stripe and prompts customers to add credit within its platform – freeing the team to focus on building.

"That time suck was preventing me from innovating on our product. Now that I don't have to manage invoicing any more, the product flywheel is turning a lot more for us, and I'm shipping more features to production," said Hodnett.

Improved visibility into customer engagement and upcoming revenue

With usage data flowing through Stripe, Chipp now has greater visibility into customer engagement beyond active sessions. Meyer can log in to the Stripe Dashboard to see token usage, model trends and how actively each customer is using the product. With a predictable 30% profit margin, Chipp can use this data to forecast revenue with precision.

"We want to track whether people are actively using our product, and when you're running agents, token usage is a better metric than simply logging in. Usage-based billing gives us visibility into how many tokens are consumed in real time. That is our one graph to rule them all," said Meyer.

With Stripe Sigma's custom reporting, Chipp's team can also dig deeper into metrics – such as revenue, transactions and churn – to make faster, data-backed decisions.

More transparency and value for clients

Chipp's token-based billing model means clients can access the best models for their needs – and have greater transparency into what they're paying for. For example, a customer can choose a faster, cheaper model for customer FAQ chatbots and a more powerful model for financial forecasting. They can see every query, every model and every cost inside Chipp, so they can make informed decisions.

"Token-based billing enables everyone to have access to the best tools. We can actually all use the same AI, and it doesn't matter if you work for a big tech company or you work for a local dental office," said Meyer.

Chipp's agency clients no longer face pricing uncertainty when pricing AI products. They can automatically track and bill their customers' token usage with a predictable markup.