Cost per model, live
Spend, tokens and cache savings resolved per model and project, with prices refreshed for you.
Every model call,
on the record.
Swap one baseURL. Get cost, latency and full agent traces for every provider you already ship on.
1 line
of config to change. The SDK call stays exactly as it is.
10
providers routed through the same endpoint today.
290s
streaming deadline, then the partial trace is still written.
MIT
licensed. Run the Docker image yourself whenever you want.
What lands in the dashboard
Five things you stop building yourself
It pages you first
Spikes, error bursts and latency drift come with the model and project that caused them.
Cost spike · gpt-4o
3.1x above the 7-day median
Pinned prompt versions
Pin a version per request, then compare cost and score on real traffic.
- support@v4live
- support@v338%
- support@v2retired
Repeat calls cost zero
One header turns on exact-match caching. Hits skip the provider and bill zero.
x-spanlens-cache: 3600412 hits today
billed at zero
Evals run inside CI
Grade with an API key, gate the merge, queue the unsure cases for a human.
$ spanlens eval run tone
38 cases · 2 flagged
score 0.91 · gate passedAgent traces
Open one trace,
see the whole run
Nested spans across tools, retrievals and model calls, with parallel branches kept intact. The slow one is obvious at a glance.
Open a sample tracetrace_9f21c4 · refund-agent
4.18 sagent.run
plan · gpt-4o
lookup_order · tool
retrieve · pgvector
policy_check · tool
answer · claude-sonnet
guardrail · eval
Keep the SDK you already import
The proxy speaks each provider natively, so streaming, tools and structured output keep working untouched.
OpenAI
chat, responses and embeddings, streaming included
Reasonable questions
- How does instrumentation work?
- Swap the provider SDK for our drop-in. Same surface, same types. We record the full request and response on the wire, with no extra round-trip and no sampling by default.
- What about latency overhead?
- p99 overhead is under 3ms. Ingestion happens async in a worker. If we ever fail, your request completes anyway. Spanlens never sits on the critical path.
- How do you handle PII?
- PII detectors (SSN, credit card, email, IBAN, passport, etc.) run at log time and flag matches for review in the Security dashboard, without blocking the request. API keys that slip into prompts are auto-masked before the row lands on disk. For workloads where prompt bodies must not be stored at all, opt out per-call with X-Spanlens-Log-Body: meta.
- Do you support OpenTelemetry?
- Yes. OTLP/HTTP ingest and export. Your existing OTel tracing flows into the same span store; LLM spans get LLM-specific attributes on top.
- What's the data retention?
- Free is 14 days. Pro is 90 days, Team is 365 days. Enterprise & self-hosted are configurable, including unlimited.
- Can I export my data?
- Anytime. JSON, CSV, Parquet. Or pipe the raw stream to S3, BigQuery, or your warehouse via our sink connectors.
- Can Spanlens tell me if a prompt actually got better?
- Yes. Evals scores responses with an LLM-as-judge on a 0 to 1 scale, per prompt version. Pair it with Experiments to replay a dataset across versions and models, so you compare quality, cost, and latency on the same inputs before you roll out.
- Does Spanlens work for a whole team?
- Projects isolate workloads, roles and invitations manage access, and audit logs record every change. Team and Enterprise add Slack, webhooks, unlimited alerts, and SSO.
Billed on requests, never on seats
Bring the whole team on any plan. The repo stays MIT, so self-hosting is always the free exit.
Free
$0/mo
For the first integration.
- 50K req / mo
- 14 day retention
- 1 seat
- All core features
- Community support
ProMost popular
$29/mo
For features already in production.
- 100K req / mo
- 90 day retention
- 3 seats
- 5 alerts · email notify
- +$8 / 100K extra
Team
$149/mo
For several products and tighter control.
- 1M req / mo
- 365 day retention
- 10 seats
- Slack · webhooks · unlimited alerts
- +$5 / 100K extra
Enterprise
Custom
For volume, SSO and a signed contract.
- Custom volume & rate limits
- 365 day retention (extendable)
- SSO (SAML / Okta)
- Unlimited seats
- Dedicated support + SLA
Self-hosted stays free. Unlimited requests, unlimited history, your own Docker image.
docker run spanlens/serverPoint one client at it today
The first request you send shows up with cost, tokens and latency attached.