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 Helicone

Zespan vs Helicone — full observability vs a proxy wrapper.

Helicone is a proxy-based LLM gateway that sits between your app and the LLM API, recording cost and latency data. It's useful for quick cost tracking but is not an observability platform — no agent identity, no delegation tracking, no evaluations, and no multi-agent support. As of mid-2025, Helicone has been reported to be in maintenance mode, with users actively seeking alternatives. Zespan offers the full picture.

Capability comparison: Zespan vs Helicone
Capability	Zespan	Helicone
Architecture	SDK + OTel ingest	Proxy (single point of failure)
Free tier	10K traces/month	100K req/month (7-day retention)
Starting price	$39/month	$25/month (Pro)
Active development	✓ Active	Maintenance mode
Multi-agent tracing	✓ Full delegation graph	Not available
Agent auto-discovery	Yes	Not available
Agent health scoring	Yes	Not available
Span waterfall	✓ Full hierarchy	Not available
Evaluations	✓ 12 templates, auto	Not available
Guardrails	✓ 7 types	Not available
Cost forecasting	✓ 30-day OLS forecast	Not available
AI cost optimizer	Yes	Not available
Natural language query	✓ ZespanPilot	Not available

Pick Zespan when…

You need real agent observability, not just cost logging.
Your system has multiple agents — delegation, handoffs, planning steps.
You want evaluations, guardrails, or runtime safety.
You can't afford a single point of failure in your LLM call path.
You want a platform that is actively developed and growing.

Pick Helicone when…

You only need basic cost and latency tracking for a single LLM endpoint.
You want a proxy-based approach with no SDK code changes at all.
You're on a legacy system where instrumenting with an SDK is not possible.

Frequently asked

What does 'maintenance mode' mean for Helicone users?

Users have reported that Helicone's feature development has slowed significantly. New agent-observability capabilities are not being added. Zespan is actively developed with new features shipping regularly.

Is a proxy architecture a problem?

Proxies add latency to every LLM call and create a single point of failure. If the proxy is down, your LLM calls fail. Zespan uses an async SDK — traces are sent in the background without blocking or adding latency to your LLM calls.

Can I migrate from Helicone to Zespan?

Yes. Remove the Helicone proxy URL and replace with Zespan's SDK wrapper (e.g., wrapOpenAI()). You get the same cost and latency data plus full agent traces, evaluations, and guardrails.

Does Zespan track cost like Helicone does?

Yes — and more. Zespan tracks cost per call, per agent, per user, per operation, and per model. It also forecasts 30-day spend and generates AI-powered recommendations to cut costs.

Try Zespan — free, no card needed

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

Get started free →