“Building billing? It sucks!”—Why Arcee built with Lago instead

Introduction

Arcee is a champion of SLMs (Small Language Models) that helps its customers build

enterprise-grade AI agents and orchestration of different AI models. To capture the demand from enterprise AI, Arcee has raised $29 million. The company is growing quickly, winning customers like Intel and constantly shipping new products.

These include Arcee Conductor, which intelligently routes prompts to the best model, and Arcee Orchestra, which allows customers to create customized AI agents using Arcee SLMs.

Like most AI products, Arcee requires more complex pricing models than a simple subscription. So when the roadmap broadened to include a second product, Arcee’s lean team decided not to reinvent billing. 

Instead, they chose Lago to handle everything from metering to invoicing.

Challenges

Arcee’s products are complex because they not only support their own models, but also work with external language models. This creates enormous billing complexity at an early stage:

• Token-level, per-model pricing: Each model needs distinct input- and output-token rates that can vary by customer plan.
  
  
• Two very different products: The billing system needs to support both usage-based API pricing as well as blended subscription/seat products. The same billing system must support a plurality of billing models (and support new ones without occupying engineers for months).
  
  
• Limited engineering bandwidth: As an early-stage startup, your number one focus is on finding product-market fit and shipping product. Building and maintaining a home-grown billing system would distract from that work.
  
  
• Self-serve and sales-assisted workflows: Sales and marketing need a UI to create special offers or tweak plans without code changes.
  
  
• Seamless payments: The team wanted to offload payments entirely to Stripe.

Solutions and impact

Lago’s biggest impact for Arcee has been not needing to waste their small team’s time building a billing system that could accommodate the flexibility they need without sacrificing functionality. 

1. Billable metrics + filters: Lago tracks input and output tokens, then filters by model so each plan inherits the right prices automatically. 
2. Instant plan management: An internal admin UI writes metadata to Arcee’s database and syncs it to Lago via API, letting non-engineers create new offers or plans in minutes.
3. Usage transparency by default: Lago’s API powers a customer-facing dashboard that breaks spend down per model and shows live wallet balances—no extra data pipeline required.

Early outcomes

• Arcee launched billing without writing their own billing code and can fully focus on their own product.
  
• Automatic invoicing and payment collection worked from day one.
  
• Non-technical team members configure bespoke deals on their own.
  
• Engineers focus solely on product code, not billing maintenance.

Monetization flexibility

As an early-stage company with complex monetization, Arcee needed both a way to implement their complex pricing now and maintain the flexibility to launch future products and charge for them. 

Lago offered exactly that with a scalable, modular approach that enables future experimentation. We’re happy to be partners with Arcee and enable them to continue iterating on pricing and monetization.