Skip to main content
Transparent infrastructure pricing

Simple, volume-based pricing

for AI unit economics.

Predictable plans built for fast-growing startups and scaling infrastructure. Choose the plan that matches your production traffic.

Bill monthlyBill annuallySave 20%

01 / Developer

Solo builders, side projects, early MVPs.

$0/month

Free forever

10K events / mo


  • 100 end users tracked
  • 7 days data retention
  • 1 team seat
  • Live cost and margin dashboard
  • Multi-provider model support
  • Basic cost attribution
  • Margin analysis
  • Threshold alerts
Start free

03 / Scale

Heavy workloads and complex agentic workflows.

$199/mo

2M events / mo

Overage: $0.10 per 1,000 events


  • 5,000 end users tracked
  • Custom data retention
  • 10 team seats
  • Full AI access
  • Priority event processing
  • Anomaly alerts
  • Budget thresholds (soft)
  • Policy enforcement
Start 14-day free trial

Enterprise

Running millions of token requests every single day?

Custom enterprise plans with high-throughput dedicated ingestion queues, custom data processing agreements (DPA), multi-year data retention, and custom SLA agreements.

Talk to the founder

No credit card required

Start the free tier instantly, no payment info.

Cancel or upgrade anytime

One click from your settings dashboard. No lock-ins.

Zero privacy risk

We read token counts, metadata, and cost endpoints. We never store raw prompts or customer data.


FAQ

Common questions, direct answers.

What exactly counts as a "Log Event"?

A log event is a single tracked AI call you send us through the SDK or API: one chat completion, one embedding request, one tool call. We read token counts, model, cost, and the metadata you attach. Your monthly limit counts these events, not tokens, and we never store raw prompts or completions.

What happens if I go over my monthly event limit?

On Growth and Scale, extra events are billed as overage ($0.15 per 1,000 on Growth, $0.10 per 1,000 on Scale), so nothing stops working. The Developer free tier caps at 10,000 events per month; when you reach it we prompt you to upgrade rather than bill you.

How long does it take to integrate Bear Lumen into my app?

Most teams are sending live data within a few minutes. Drop in the Node.js or Python SDK, or wire up the MCP and Skill files, and your cost and margin dashboard starts filling in immediately. No data warehouse or pipeline work required.

Do you offer custom pricing if we bring our own self-hosted models?

Yes. Enterprise plans support self-hosted and custom models, dedicated ingestion queues, a data processing agreement (DPA), multi-year retention, and a custom SLA. Talk to the founder and we will shape a plan around your stack.