Zespan is an AI agent observability and engineering platform. It traces every agent decision, tool call, handoff, and delegation in production. It also provides prompt versioning, built-in LLM-as-judge evaluations, guardrails, cost optimization, and an AI ops assistant called ZespanPilot.

How do I instrument my AI agent with Zespan?

Zespan requires 2 lines of code. Import zespan and call zespan.init({ apiKey: process.env.LT_KEY }). This auto-patches OpenAI, Anthropic, Gemini, Bedrock, and Mistral. For framework-level tracing, add one handler: ZespanCallbackHandler for LangChain, ZespanCrewAIListener for CrewAI, or ZespanADKHandler for Google ADK.

Does Zespan support prompt versioning?

Yes. Zespan includes prompt management with versioning, a playground for iteration, and A/B testing to compare prompt versions against each other in production.

What evaluations does Zespan support?

Zespan ships 12 built-in LLM-as-judge evaluation templates including faithfulness, relevance, toxicity, groundedness, and more. Evaluations run automatically on every trace with no custom scoring functions required.

How does Zespan compare to Langfuse?

Zespan is agent-native: every span carries agent identity, delegations are first-class trace events, and an agent map is built automatically. Langfuse was built for LLM pipelines and extended to agents later. Zespan also ships 12 built-in eval templates (Langfuse has none), includes an AI cost optimizer, and ZespanPilot for AI ops. Langfuse has open-source self-hosting; Zespan does not.

What is the free tier for Zespan?

The free tier includes 10,000 traces per month, 14-day retention, 2 projects, and 1 seat. No credit card required.

Compare

Zespan vs LangSmith

Zespan vs LangSmith — agent-first vs chain-first observability.

LangSmith is the best tool if your entire stack is LangChain/LangGraph and your team is comfortable with its per-seat pricing. Outside that ecosystem it becomes friction-heavy — no auto-discovery, poor UX for CrewAI or AutoGen agents, and no AI-powered cost optimization. Zespan is built agent-first and works across every major framework from day one.

Capability comparison: Zespan vs LangSmith
Capability	Zespan	LangSmith
Free tier	10K traces/month	5K traces/month
Starting price	$39/month	$39/seat/month
Framework support	Any framework + OTel	Best for LangChain/LangGraph
Agent auto-discovery	✓ 2 lines, zero config	Not available
Multi-agent delegation graph	Yes	Limited (LangGraph only)
Agent health scoring	✓ A–F composite score	Not available
LLM-as-judge auto-evals	✓ 12 templates, automatic	Manual only
AI cost optimizer	✓ 30-day forecast + recs	Not available
Natural language query	✓ ZespanPilot NLQ	Not available
Guardrails	✓ 7 types, pre/post	Not available
Incident management	✓ Full state machine	Not available
OpenTelemetry native	✓ OTLP/HTTP endpoint	Partial

Pick Zespan when…

Your agents span multiple frameworks — CrewAI, AutoGen, Google ADK, custom.
You want cost forecasting, AI optimization recommendations, and anomaly detection.
You need guardrails, PII redaction, or runtime safety in production.
You want an AI copilot (ZespanPilot) for querying observability data in plain English.
You're evaluating at the team/org level and don't want per-seat pricing.

Pick LangSmith when…

Your stack is 100% LangChain or LangGraph and you want the deepest native integration.
You want LangGraph state diffs between graph nodes — LangSmith's unique differentiator.
Your team already has LangSmith in its workflow and switching cost outweighs the gaps.

Frequently asked

Does Zespan support LangChain?

Yes. Zespan ships a drop-in LumiqCallbackHandler for LangChain and LangChain.js. It traces every chain step, agent reasoning, and tool call — the same way LangSmith does, but with the full Zespan feature set on top.

How does Zespan handle multi-framework teams?

Zespan supports LangChain, CrewAI, AutoGen, Google ADK, LlamaIndex, Vercel AI SDK, and PydanticAI out of the box — plus any OpenTelemetry-instrumented framework via the native OTLP/HTTP endpoint. One dashboard for your whole agent stack.

Is Zespan more expensive than LangSmith?

LangSmith starts at $39/seat/month. Zespan starts at $39/month flat for the Solo plan (one seat). For teams of 3+, Zespan's Pro plan at $149/month (10 seats) is typically cheaper than LangSmith's per-seat model.

Can I migrate from LangSmith to Zespan?

Yes. Replace LangSmith's callback or OTEL exporter with Zespan's drop-in equivalent. For LangChain apps, swap LangSmithTracer for LumiqCallbackHandler. Setup takes under 5 minutes.

Try Zespan — free, no card needed

10K traces/month free. Setup takes under 5 minutes. See why teams switch from LangSmith.

Get started free →