Your AI works in demos. Make it work in production.
TruLayer is the reliability platform for AI agents — combining real-time tracing, automated evaluation, and self-healing controls so your agents stay on-rails at scale.
Free through V1-P1 GA for design partners · No credit card
Works with
Capabilities
Beyond observability
Observability shows you the problem. Reliability means you've already solved it. TruLayer closes the loop — detect, evaluate, and remediate automatically.
Agent Step Tracing
Capture every tool call, LLM hop, and decision branch across multi-agent workflows — not just the final response.
Intermediate Evals
Score outputs at each step, not just the end. Catch hallucinations and off-rail reasoning mid-chain before they compound.
Real-time Alerts
Configurable rules fire instantly when error rates, latency, or cost exceed thresholds. Email and webhook delivery.
Auto Remediation
Close the control loop — TruLayer can retry with a fallback model or modified prompt when an agent misbehaves.
Session Replay
Step through an agent run chronologically. See the full reasoning chain with latency breakdowns and token counts.
Multi-tenant RBAC
Isolate data per project and org. Owner, member, and viewer roles. Enterprise SAML via Clerk with zero config.
Instrument in minutes
Two lines. Full visibility.
Wrap any OpenAI, Anthropic, or custom LLM client with TruLayer. Every call, chain, and tool use is automatically traced — no manual spans, no config files.
- ✓Auto-captures inputs, outputs, tokens, latency, cost
- ✓Propagates trace context across async agent hops
- ✓Zero overhead — proxy-based, not monkey-patching
import TruLayer from 'trulayer'
import OpenAI from 'openai'
const tl = new TruLayer({ apiKey: process.env.TRULAYER_API_KEY })
const openai = tl.instrument(new OpenAI())
// Every call is now traced, evaluated, and monitored
const response = await openai.chat.completions.create({
model: 'gpt-4o',
messages: [{ role: 'user', content: task }],
tools: agentTools,
})
// TruLayer automatically captures:
// ✓ input / output / tool calls
// ✓ token usage and cost
// ✓ latency per hop
// ✓ parent trace across async agent stepsHow it works
From deploy to confidence in three steps
Instrument
Wrap your LLM client with two lines of code. Supports OpenAI, Anthropic, and any custom model.
Observe
Every agent step, tool call, and chain hop appears in your trace explorer in real time.
Evaluate & Fix
Automatic evals score every output. Failure alerts and auto-remediation close the loop.
Pricing
Simple, usage-aligned pricing
Pilot access is free for design partners through V1-P1 GA. The numbers below are our published GA hypothesis — shared now so you can tell your manager what this will cost. They are not a rate lock.
Starter
Self-serve, no credit card
- 500K spans / month
- OSS eval catalog
- Community support
Pro
Hypothesis at V1-P2 GA
- 5M included spans
- 50K eval runs / month
- $40 / additional 1M spans
- Email support
Team
Hypothesis at V1-P2 GA
- 20M included spans
- $35 / additional 1M spans
- SSO + RBAC
- Shared Slack channel
Enterprise
Committed-use contract
- Negotiated span volume
- Overage caps
- DPA + custom terms
- Dedicated support
How the pilot converts. Design partners get full Pro-tier access free through V1-P1 GA (target 2026 Q2 end). At V1-P1 GA you receive a 30-day notice email with the full V1-P2 pricing schedule attached — no auto-conversion, no silent billing. You opt into a paid plan or walk away with no obligation. The numbers above are our working hypothesis; actual V1-P2 pricing may move based on pilot usage data. We do not offer rate locks or grandfathered pricing.
Reliable AI.
Not just observable AI.
Observability tells you what broke. TruLayer tells you what broke, why, and fixes it automatically. Pilot access is free for design partners through V1-P1 GA.
Join the pilot