Last updated 06-28-2026
Category:
Reviews:
Join thousands of AI enthusiasts in the World of AI!
Galileo
Galileo is an AI observability and eval engineering platform for teams shipping LLM applications and agents. It helps you evaluate models offline, monitor production traffic, and turn passing evals into real-time guardrails so failures get caught before users do. The platform targets developers and enterprise AI teams who need measurement across the full build-to-production cycle.
Galileo distills expensive LLM-as-judge evaluators into compact Luna models that run with low latency and lower cost, making it practical to score 100% of production traffic. The insights engine analyzes agent behavior to surface failure modes and suggest fixes, while 20+ built-in evaluators cover RAG, agents, safety, and security out of the box. Cisco has announced its intent to acquire Galileo.
Founded in 2021, Galileo was built around the idea that non-deterministic LLMs need reliability workflows spanning evaluation, monitoring, observability, and protection. The founders drew on experience from Google AI, Uber AI, and Apple Siri. Customers can deploy via SaaS, virtual private cloud, or on-premises depending on security needs.
Galileo fits ML engineers, platform teams, and AI product leads who need datasets, custom metrics, traces, and guardrail policies in one place instead of stitching together separate testing and monitoring tools.
20+ built-in evaluators for RAG, agents, safety, and security
Distill LLM-as-judge evaluators into Luna models at 96% lower cost
Luna-2 evaluators average under 200ms for real-time production guardrails
Insights engine pinpoints agent failure modes and suggests concrete fixes
Auto-tune custom metrics with continuous learning from human feedback
Deploy on SaaS, VPC, or on-premises with SDK and API integrations
Covers offline evals, production observability, and real-time guardrails in one platform.
Free tier includes 5,000 monthly traces with unlimited users and custom evals.
Luna models make full-traffic production evaluation practical at lower latency and cost.
Enterprise deployment options include VPC, on-premises, SSO, and dedicated inference servers.
Pro pricing scales with trace volume beyond the listed 50,000-trace allowance.
Enterprise features like SSO, on-prem deployment, and phone support require a sales conversation.
Advanced guardrails and unlimited traces are not available on the Free plan.
Does Galileo have a free plan?
Yes. Galileo offers a Free plan at $0 per month with 5,000 traces per month, unlimited users, and unlimited custom evals. You can sign up at app.galileo.ai without booking a sales call.
What does Galileo Pro include?
Galileo Pro costs $100 per month when billed annually and includes everything in Free plus 50,000 traces per month, standard RBAC, advanced analytics and insights, and dedicated Slack support. Pricing scales based on trace volume.
What deployment options does Galileo support?
Galileo supports SaaS, virtual private cloud, and on-premises deployment. Enterprise plans add custom rate limits, enterprise-grade RBAC, SSO, real-time guardrails, and optional low-latency dedicated inference servers.
What are Luna models in Galileo?
Luna models are Galileo's compact evaluation models distilled from larger LLM-as-judge evaluators. Luna-2 variants run with sub-200ms average latency so teams can monitor production AI traffic in near real time at lower cost than full judge models.
What types of AI systems can Galileo evaluate?
Galileo supports evaluation and observability for RAG pipelines, multi-step agents, chatbots, and general GenAI applications. The platform includes prebuilt metrics plus tools to create custom code-based or LLM-as-judge evaluators.
How do I contact Galileo support?
You can reach Galileo at [email protected] for privacy and general questions. Pro customers get dedicated Slack support, and Enterprise plans add 24/7 support via Slack, email, or phone plus a dedicated customer success manager.