Galileo pricing: what’s included in Free (5,000 traces/month) vs Pro ($100/month, 50,000 traces/month), and how do overages work?

Most teams don’t hit AI reliability problems because of a missing feature—they hit them because they underestimate how much evaluation and observability traffic they’ll actually generate. Galileo’s Free and Pro plans are designed around that reality, with traces as the core unit and pricing that scales with how much you’re actually testing and protecting in production.

Quick Answer: Galileo’s Free plan gives you 5,000 traces/month, unlimited users, and unlimited custom evals so you can stand up evals and basic observability at no cost. Pro starts at $100/month for 50,000 traces/month plus RBAC, advanced analytics, and dedicated Slack support. Pricing scales with trace volume; for higher or more variable usage (including overages at scale), teams typically step up to higher Pro tiers or Enterprise.

---

The Quick Overview

• What It Is: A trace-based pricing model for Galileo’s AI reliability platform, with Free and Pro tiers that scale by how many evaluation/observability traces you run each month.
• Who It Is For: Builders shipping LLM apps, RAG systems, and agents who need to move from “demo-grade” reliability to production guardrails while controlling cost.
• Core Problem Solved: You get a predictable way to budget for evaluation and guardrails—so you can run continuous evals in development and production without getting surprised by an opaque “AI monitoring” bill.

---

How It Works

Galileo prices around traces per month, not seats or generic “projects.” A trace is a full session or workflow run—spans, tool calls, prompts, and model responses—captured and evaluated by Galileo across Evaluate, Signals, and Protect.

Here’s the mental model:

• Each time your app sends a request (or runs an agent workflow) and you log it to Galileo, you consume trace capacity.
• Evals (whether out-of-the-box or custom) run against those traces.
• As you ramp from small experiments → staging → production, you move from the Free tier to Pro and, for unlimited scale, to Enterprise.

1. Free Plan (5,000 traces/month):
   Designed for early-stage builds and small teams. You get full access to core eval workflows—unlimited custom evals, 5,000 traces/month, and unlimited users—so you can define your reliability requirements and run real evals before you ever pay.

2. Pro Plan ($100/month, 50,000 traces/month):
   Built for teams moving into production and needing more traffic coverage, permissions, and deeper insights. You get 10x the traces, plus Standard RBAC, advanced analytics & insights, and dedicated Slack support. Pricing scales with trace volume—if you outgrow 50,000 traces/month, you can move to a higher Pro tier or Enterprise.

3. Enterprise Plan (Contact Us, Unlimited traces):
   For organizations that need unlimited trace capacity, custom rate limits, and real-time guardrails at scale. You get hosted/VPC/on‑prem deployment options, enterprise RBAC and SSO, dedicated CSM, 24/7 support, low-latency dedicated inference servers, and forward-deployed engineering support.

---

Features & Benefits Breakdown

Core Feature	What It Does	Primary Benefit
Trace-based pricing	Meters usage by traces (sessions/workflows) instead of seats or projects.	Aligns cost with real evaluation and guardrail coverage; easy to forecast as traffic grows.
Unlimited custom evals	Lets you define any number of evaluators tailored to your domain.	You’re never penalized for designing better evals; complexity doesn’t inflate your bill.
Plan-specific capabilities	Adds RBAC, analytics, support, and deployment options as you move up tiers.	You can start free, then layer in governance, visibility, and support as you go to production.

---

What’s Included in the Free Plan (5,000 traces/month)

The Free plan exists so you can build a serious eval engineering workflow without pulling out a credit card.

Included in Free:

• 5,000 traces per month
  • Enough to instrument early agent/RAG prototypes, run iterative prompt/model tests, and capture initial production traffic.
• Unlimited users
  • Bring in PMs, QA, and domain experts to review traces and annotate issues without seat friction.
• Unlimited custom evals
  • Design as many evaluators as you need:
    • Hallucination checks for your specific knowledge base.
    • Safety and policy adherence for your domain.
    • Tool selection quality and tool-call success for agents.
• Core Evaluate + Signals experience
  • Build golden test sets from dev and synthetic data.
  • Analyze early failure modes across your traces.

Best fit for:

• Teams still in prototype or early staging.
• One or a few AI services with low-to-moderate traffic.
• Org-wide exploration where different squads want to try Galileo without cost friction.

---

What’s Included in the Pro Plan ($100/month, 50,000 traces/month)

Pro is where you go when “this is a cool demo” turns into “we’re shipping this to customers.”

Everything in Free, plus:

• 50,000 traces per month (starting)
  • 10x Free capacity, suitable for:
    • Production RAG/agent systems with consistent traffic.
    • Wider experimentation across models, prompts, and tools.
• Standard RBAC
  • Control who can:
    • Create/edit evaluators and guardrail policies.
    • Access sensitive traces (PII, internal tools).
    • Manage deployments and environment configs.
• Advanced analytics & insights
  • Deeper visibility into:
    • Failure patterns across sessions → traces → spans.
    • Latency and cost per trace, per route, per model.
    • Drift and degradation over time.
• Dedicated support via Slack
  • Direct access to the Galileo team:
    • Help designing effective evals for your domain.
    • Guidance on routing, guardrails, and instrumentation.
    • Faster resolution when something looks off in production.

How Pro scales and how overages fit in:

• The entry Pro tier is $100/month, which includes 50,000 traces/month.
• As your traffic grows beyond 50,000 traces/month, pricing scales based on trace volume.
  • Practically, that means:
    • You can work with Galileo to size the right Pro tier for your expected traffic.
    • If you’re consistently over 50,000 traces/month, you’ll graduate into a higher-volume Pro tier or Enterprise.
• For very high or highly variable traffic (e.g., spikes to millions of traces/month), teams typically switch to Enterprise to get unlimited traces and custom rate limits.

If you’re forecasting rapid growth or are already running multiple high-traffic agents, assume that Pro is your starting point, not your ceiling.

---

How Overages Work in Practice

Galileo’s pricing is set up to avoid surprises, especially when you’re treating evals as always-on guardrails instead of a one-off test run.

Here’s how to think about overages and growth:

• Free → Pro transition:
  • When your usage regularly approaches 5,000 traces/month, that’s the signal you’re outgrowing Free.
  • You’ll move to Pro to get 10x trace capacity and better governance/analytics.
• Within Pro:
  • Pro starts at 50,000 traces/month. Pricing then scales with trace volume.
  • If you’re briefly over 50,000 in a given month, the Galileo sales team can help size a higher-volume Pro configuration so you’re not repeatedly bumping the ceiling.
• Beyond Pro:
  • If you need effectively unlimited traces, or your reliability posture depends on 100% coverage across several products, you’ll move to Enterprise, which includes:
    • Unlimited traces.
    • Custom rate limits.
    • Low-latency dedicated inference servers.
    • Real-time guardrails (Protect) and always-on monitoring at scale.

The key: Galileo is explicit that pricing scales based on number of traces. As soon as your usage pattern looks more like “we’re guarding 24/7 production traffic” than “we’re running a limited pilot,” you’ll align with a plan size that matches that footprint rather than relying on ad hoc overages.

---

When to Choose Free vs Pro

• Stick with Free if:

  • You’re still instrumenting your first RAG or agent workflow.
  • Your traffic volume is modest (e.g., internal dogfooding, pre-launch beta).
  • You want to explore Galileo’s Evaluation Engine, Signals, and custom eval design without procurement overhead.

• Move to Pro if:

  • You’re sending external users to an AI feature and want consistent reliability.
  • You’re evaluating multiple model/prompt combinations and need 50,000+ traces/month.
  • You need RBAC, richer analytics, and dedicated Slack support to manage multiple teams or services.

For orgs with strict security requirements, 24/7 SLAs, or multi-region deployments, the right choice is typically Enterprise, where you add deployment flexibility (SaaS, VPC, on‑prem), SSO, and low-latency dedicated inference.

---

Pricing & Plans Overview

From the official Galileo pricing:

• Free — $0/month

  • 5,000 traces per month
  • Unlimited users
  • Unlimited custom evals

• Pro — $100/month*

  • 50,000 traces per month
  • Standard RBAC
  • Advanced analytics & insights
  • Dedicated support (Slack)
  • Pricing scales based on number of traces.

• Enterprise — Contact us

  • Everything in Pro, plus:
    • Unlimited traces
    • Custom rate limits
    • Deploy: Hosted, VPC, or on‑prem
    • Enterprise-grade security RBAC, SSO
    • Dedicated CSM
    • Real-time guardrails
    • 24/7 support: Slack, email, or phone
    • Low-latency dedicated inference servers
    • Forward deployed engineering support

---

Frequently Asked Questions

Can I really run unlimited custom evals on the Free plan?

Short Answer: Yes. Both Free and Pro let you create unlimited custom evals.

Details:
Galileo doesn’t charge you per evaluator or per metric. You can define as many custom evals as you need—for hallucinations, prompt injection, PII exposure, policy violations, tool call correctness, and more. What’s metered is the number of traces you run those evals on. That means you’re free to iterate on evaluation design without watching the meter on eval count.

---

What happens if I outgrow 50,000 traces/month on Pro?

Short Answer: Pricing scales with your trace volume; you’ll move to a higher-volume Pro configuration or Enterprise.

Details:
The published Pro starting tier is $100/month for 50,000 traces/month. If your app’s reliability strategy depends on evaluating more traffic—say, you need to cover all agent sessions across multiple products—Galileo will size a Pro or Enterprise configuration that fits your expected monthly traces. Enterprise is typically where teams land once they need unlimited traces, custom rate limits, and dedicated infrastructure (like low-latency Luna-2 inference) for real-time guardrails.

---

Summary

Galileo’s pricing is built around a simple idea: traces are the real unit of AI reliability. Free gives you 5,000 traces/month plus unlimited users and custom evals so you can design and validate your evaluation strategy. Pro starts at $100/month for 50,000 traces/month and adds RBAC, advanced analytics, and dedicated support as you scale into production. As your traffic grows, pricing scales with trace volume, and Enterprise unlocks unlimited traces and full enterprise deployment, security, and support.

If you’re serious about turning offline evals into always-on guardrails, plan for the point where you move from Free experiments to Pro (or Enterprise) production coverage—because reliability comes from running your best evals on real traffic, continuously.

---

Next Step

Get Started