Charge by ingest, the way logs are priced. Add the whole company at no extra cost. Free up to 50,000 spans a month — and never any surprise cardinality bills, no matter how many tags you set.
For solo builders, side projects, and the first weekend of a new agent. Forever free, no credit card.
For teams shipping LLM features in production. Linear pricing on ingest, no per-seat tax, no surprise cardinality bills.
For regulated teams and 100+ engineer orgs. Self-host or deploy in your VPC. Volume pricing per ingest tier.
Drag the sliders. Most teams settle into the Pro tier somewhere between 5M and 25M spans a month. Pricing is linear past the included bucket.
Three real workload profiles from current customers, side-by-side with the closest alternative we benchmarked. Numbers are public list pricing as of April 2026.
A span is one unit of work in your trace — an LLM call, a tool call, a retry, a custom span you wrap with our SDK. A typical agent turn produces 5–15 spans. We don't charge for trace metadata, dashboards, or seats — only for span ingest volume.
On Pro you're billed at $0.10 per 1,000 additional spans. There is no hard cutoff — your traces continue ingesting and you'll see the overage on your next monthly invoice. You can set a budget cap that throttles new ingest if you'd prefer hard limits.
Because LLM observability is a "the whole team needs to look at this" tool, not "the SRE on call." Per-seat pricing punishes teams for letting more engineers see their own production telemetry. We charge for ingest because that's what costs us money.
Yes — eval runs are themselves spans, and they're billed at the same rate as any other span. Most teams find evals add roughly 10–15% to their span count when running the recommended five-panel default.
Yes. Companies under $5M ARR get 50% off Pro for the first year, no strings, no equity ask. Email startups@yourbusiness.com with a one-line description and we'll send a code within a day.
Yes, on Enterprise. We ship a Helm chart for Kubernetes and a Terraform module for AWS, GCP, and Azure. Most self-hosted deploys are running within two weeks of contract signature.
If you cancel within the first 30 days of a Pro plan, we refund the full month, prorated against any over-bucket usage. After that, we don't offer prorated refunds, but you can downgrade to free at any time and keep your data for 30 days.
No credit card to start. Drop in three lines of SDK and you'll see your first trace in the dashboard in under a minute.