> ## Documentation Index
> Fetch the complete documentation index at: https://docs.zespan.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Zespan: AI Agent Reliability Platform
> Zespan doesn't just watch your AI agents — it blocks unsafe runs mid-execution from inside your own process, classifies every trace pass/fail with a deterministic verdict, and clusters recurring failures automatically. Engineering teams ship agents they can trust, not just agents they can watch fail.
Zespan is an **AI agent reliability platform** built for teams shipping AI applications powered by agents. Watching your agents is table stakes — Zespan is built to keep them working. Whether you're running a single-step RAG pipeline, a multi-agent coordinator with specialists, or a complex agentic workflow with tool use and handoffs, Zespan captures every step at runtime, enforces the policies that keep a run from going wrong while it's still running, and tells you — with a real verdict, not a wall of raw scores — whether the result was actually good.
Enforce block mid-run
Verdict pass / fail, deterministic
Cluster group recurring failures
## What makes an agent reliable
Block an agent mid-run on loop, scope, budget, or delegation policy — a library call inside your own process, no proxy, no separate gateway contract.
Every trace arrives pre-classified — response, trajectory, or step level — so you know whether a run actually succeeded, not just what it cost or how long it took.
The same deterministic verdict system groups repeated failures into a single tracked Issue, so you stop re-discovering the same bug one trace at a time.
Follow a user request across every agent turn, tool invocation, sub-agent delegation, and LLM call. The flame graph shows each span's latency, cost, and status.
See coordinator-to-specialist handoffs, parallel agent runs, and delegation chains, plus a cross-project trust ledger on every agent's guardrail-compliance history.
Token costs attributed per agent, per tool, per model, per user, and per session — plus a cost-vs-quality frontier that flags a cheaper model scoring just as well.
## Who Zespan is for
Zespan is built for engineers and teams building production AI applications with agents. If your application makes LLM calls, uses tools, orchestrates multiple agents, or runs RAG pipelines, you need more than visibility into what happened — you need the ability to stop a run before it does damage, and a straight answer on whether it succeeded. Zespan gives you both, with two lines of SDK setup.
You don't need to change how you build your agents. Wrap your existing LLM clients and frameworks, and Zespan instruments everything automatically — including LangChain chains, Google ADK agents, and custom multi-agent systems.
## Pricing
Zespan charges per organization. Start on the Free tier with no credit card required, then scale up through Solo, Pro, and Team plans as your agent traffic grows, or talk to us about a custom Scale plan for high-volume and enterprise needs.
See the full plan comparison, trace quotas, and overage rates on the Billing page.
## Ready to start?
Instrument your first agent and see it in the dashboard in under five minutes.
Install the SDK, wrap your agent or LLM client, and see your first trace appear in the dashboard.
