Both cover agent tracing, offline experiments, online evaluation and human annotation. Galileo, now part of Splunk, adds guardrails that act on model and tool inputs and outputs during a run; HoneyHive is an independent, OpenTelemetry-native platform with prompt versioning and a free tier. Whether you need enforcement from the same vendor, and which buying route suits you, usually decides it.[3][4]
Galileo and HoneyHive both trace and evaluate agents, but Galileo can also block harmful model and tool inputs and outputs during a run, and HoneyHive only observes and scores. Galileo has been sold as Splunk Agent Observability since 7 August 2026, after Cisco completed its acquisition in May. HoneyHive is a small independent company with a free Developer tier and quote-based Enterprise pricing, and no published self-serve paid plan.
| Decision factor | Galileo | HoneyHive |
|---|---|---|
| Ownership | Cisco, completed May 2026; sold as Splunk Agent Observability since 7 August 2026 | Independent; $7.4M total funding announced April 2025 |
| Run-time guardrails | Yes: Agent Control checks model and tool inputs and outputs during a run | None; observes and scores, and logs results from your own guardrails |
| OpenTelemetry | Accepts OTLP; OpenInference instrumentors in Python only | Built on it; OTel GenAI, OpenLLMetry and OpenInference conventions |
| Prompt tooling | Prompts and a playground; version control not documented | Playground, prompt versioning and deployment by name or version |
| Evaluators | LLM judges, Python code metrics and Luna-2 small models; SaaS edition limited to five preset Luna evaluators | LLM evaluators with custom rubrics, server-side Python evaluators and human review |
| Self-hosting | On-premises, including air-gapped (commercial) | Enterprise only, fully self-hosted or hybrid, documented on AWS |
| Free tier | None listed for Agent Observability | Developer: 10,000 events a month, 30-day retention, up to 5 users |
| Paid pricing | From $100 a month per 1.2 million spans a year, billed annually | No published self-serve plan; Enterprise is quote-based |
| Capability | Galileo | HoneyHive | Prefactor |
|---|---|---|---|
| Observe | |||
| LLM call tracing | ✓ | ✓ | ✓ |
| Multi-step agent tracing (tool calls, nested spans) | ✓ | ✓ | ✓ |
| Cost tracking | ✓ | ✓ | ✓ |
| OpenTelemetry support | ✓ | ✓ | ✓ |
| Evaluate | |||
| Offline evals (datasets, experiments) | ✓ | ✓ | ✓ |
| Online evals on production traffic | ✓ | ✓ | ✓ |
| LLM-as-a-judge scoring | ✓ | ✓ | ✓ |
| Code-based evaluators | ✓ | ✓ | ✓ |
| Human review and annotation | ✓ | ✓ | ✓ |
| Prompt versioning and playground | Partial | ✓ | — |
| Act on findings | |||
| Runtime controls on live agent actions | ✓ | — | ✓ |
| Gateway or proxy in the request path | — | — | — |
| Licence and deployment | |||
| Open source | Partial | Partial | Partial |
| Self-hosting | ✓ | ✓ | Partial |
This is the main product difference. Galileo's Agent Control evaluates model and tool inputs and outputs during agent execution and can block prompt injection, personal data leaks and harmful content using rules set in one place, without changes to agent code. It replaced the older Protect product in June 2026, and its core is open source under Apache-2.0. Blocking content is documented; approval workflows for actions are not.
HoneyHive does not document blocking, guardrails or approvals on live agent actions. Its Policies feature sets organisation-wide evaluation and monitoring standards, not run-time enforcement. In Microsoft's ASSERT and Agent Control Specification announcement in June 2026, HoneyHive was named as the observability partner, with enforcement handled by the specification itself. If you already enforce in your own code or another layer, HoneyHive can log those guardrail results as custom metrics.
HoneyHive is built on OpenTelemetry and accepts traces using the OTel GenAI, OpenLLMetry and OpenInference conventions, with a Collector integration. It offers agent graphs, trajectories, threads and, since September 2026, a conversation view, plus tracing for coding agents such as Claude Code, Codex CLI and Devin. Experiments can gate regressions in CI through GitHub Actions, and deployed prompts are fetched by name or version from code. Its full tracing SDK is Python; the TypeScript package is a REST client, so TypeScript teams may lean on OTLP.
Galileo accepts OTLP and supports OpenInference instrumentors in Python only, alongside its Python and TypeScript SDKs and native integrations for agent frameworks and A2A tracing. It adds Luna-2 small models as lower-cost judges and multi-turn experiments since June 2026. The Splunk-hosted edition supports five preset Luna evaluators and no custom code evaluators; those are on-premises only. Its prompts page does not document version control, so teams that version prompts heavily will find more in HoneyHive.
Neither runs its own gateway for application traffic. HoneyHive integrates with third-party gateways such as LiteLLM, Portkey and Cloudflare AI Gateway, and its evaluator calls can be routed through Portkey. Galileo ingests through SDKs, integrations, OTLP and its API.
Galileo now sits inside Cisco's Splunk, with annual billing and a choice of SaaS or on-premises deployment, including air-gapped. Customers onboarded before 7 August 2026 use Galileo's documentation; later customers use Splunk's.
HoneyHive is a small company: its last announced funding was $7.4M in total in April 2025, and third-party directories list around 18 to 19 staff. It is SOC 2 Type II and offers a HIPAA BAA on Enterprise. The buying routes differ: HoneyHive lets you start free and requires a sales conversation only beyond the Developer tier, while Galileo has no free starting point and bills annually from the outset.
Deployment options differ in detail. HoneyHive offers multi-tenant SaaS in AWS us-west-2, a dedicated single-tenant cloud, a hybrid model with your data plane and its control plane, and full self-hosting in your own AWS account or VPC on Enterprise. Splunk Agent Observability runs as SaaS in Splunk Observability Cloud or on-premises, including air-gapped. If you must self-host outside AWS, Galileo documents that path and HoneyHive does not.
As of 20 October 2026, HoneyHive's Developer tier is free with no card, 10,000 events a month, 30-day retention and up to five users. There is no published self-serve paid plan; Enterprise, including self-hosting, is quote-based. Splunk lists Agent Observability from $100 a month per 1.2 million spans a year, billed annually, with a Luna token allowance and no free tier listed.
The free tier has other limits worth checking before you rely on it: one workspace and 1,000 requests a minute, on multi-tenant SaaS in a single AWS region. Splunk lists extended retention and extra Luna tokens as paid options.
Neither publishes a self-serve monthly plan above its entry point. Read Splunk's pricing and HoneyHive's pricing, then ask both for a quote at your expected volume.
We sell the layer this section describes. Read it with that in mind.
Stated plainly, since Prefactor publishes this page: Prefactor is an agent evaluation platform that scores each step of a run and acts on what it finds through runtime policies on tool calls and data access, approval holds for risky actions, and a kill switch for a run, an agent or a team. Galileo's Agent Control is the closer of these two, since it blocks content during a run; HoneyHive observes and scores. Neither documents approval holds. Prefactor has no prompt playground or prompt registry and its platform is hosted, so if prompt versioning is central to your workflow, HoneyHive covers it and Prefactor does not. See the full comparison of 15 platforms.
Reviewed against public product and documentation pages on October 20, 2026. If a vendor has changed a feature, product name, or positioning since then, send a correction and we will update it. Numbered source links in the page body point to the ordered sources below.
Prefactor scores each step of every agent run and can hold or stop an action when a score says it should not happen. Start on the free tier: 5,000 spans a month, no credit card required.
Prefactor helps teams observe, evaluate, and improve their AI agents in production, across every framework and provider.