Both products are now called Agent Observability, and both live inside a wider monitoring suite. Pick by suite first: Datadog for Datadog estates, Galileo for Splunk and Cisco estates. Galileo includes run-time blocking and on-premises deployment; Datadog includes a free tier and versioned prompts.[1][6]
Both are now modules of large observability suites: Datadog Agent Observability, formerly LLM Observability, sits in Datadog, and Galileo has been sold as Splunk Agent Observability since 7 August 2026, after Cisco completed its acquisition in May 2026. Galileo's Agent Control can block harmful model and tool inputs and outputs during a run; Datadog's module does not block, and sells that separately as AI Guard. Galileo can run on premises; Datadog is hosted only.
| Decision factor | Datadog | Galileo |
|---|---|---|
| Ownership | Datadog, a public company (Nasdaq: DDOG); needs a Datadog account | Acquired by Cisco (completed May 2026); sold as Splunk Agent Observability since 7 August 2026 |
| Run-time blocking | Not in the module; AI Guard is a separate product in Limited Availability | Yes: Agent Control checks model and tool inputs and outputs during a run |
| Evaluators | LLM-as-a-judge and built-in evaluators; code evaluators automatic only in Experiments | LLM judges and Luna-2 small models; custom code metrics on premises only |
| Prompt tooling | Versioned prompt registry, per-environment deployment, A/B tests | Prompts and a playground; version control not documented |
| Self-hosting | None documented | On premises, including air-gapped (commercial) |
| Free tier | Free: 40K LLM spans a month, 15-day retention | None listed for Agent Observability |
| Paid pricing | Pro: $160 a month billed annually, 100K LLM spans | From $100 a month per 1.2 million spans a year, billed annually |
| Capability | Datadog | Galileo | 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 | Partial | ✓ | ✓ |
| Gateway or proxy in the request path | — | — | — |
| Licence and deployment | |||
| Open source | Partial | Partial | Partial |
| Self-hosting | — | ✓ | Partial |
Datadog renamed LLM Observability to Agent Observability in June 2026. Cisco completed its acquisition of Galileo in May 2026, and since 7 August 2026 Galileo has been sold as Splunk Agent Observability inside Splunk Observability Cloud. Customers who joined Galileo before that date still use Galileo's documentation and app; later customers use Splunk's.
Both are now sold as parts of a wider suite. Datadog's module needs a Datadog account and links agent traces to Datadog APM and infrastructure data. Galileo's value grows if infrastructure and application telemetry already go to Splunk. For most buyers the suite they already run decides this before any feature comparison. If you run both, or neither, the rest of this page applies.
Galileo's Agent Control evaluates model and tool inputs and outputs while an agent runs and can block prompt injection, personal data leaks and harmful content, using rules set in one place and without changes to agent code. It replaced the older Protect product in June 2026, and its core was released as open source under Apache-2.0 in March 2026. The hosted version is documented as an enterprise feature. It blocks content; approval workflows for actions are not documented. Agent Control's open-source core can also be self-hosted from its own repository, separately from the rest of the platform.
Datadog Agent Observability monitors and evaluates but does not block. Blocking prompts, responses and tool calls is offered by AI Guard, a separate Datadog product in Limited Availability as of mid-2026 and billed per token scanned. If run-time checks are a requirement this quarter, Galileo includes them; with Datadog you are buying a second product with limited availability.
Both run experiments on datasets, score sampled production traffic with LLM judges and offer annotation queues for human review. Galileo adds Luna-2 small models as lower-cost judges and multi-turn experiments. One limit to check: the Splunk-hosted edition supports five preset Luna evaluators and no custom code evaluators, which are available on premises only. Datadog's code evaluators run automatically in Experiments; for production traffic you run them in your application and submit the scores.
On prompts, Datadog has the stronger documented tooling: a versioned registry retrieved at run time from Python, Go or JavaScript, per-environment deployment with targeting rules, and A/B tests of prompt versions. Galileo stores and templates prompts and has a playground, but its prompts page does not document version control.
For agent tracing, Galileo has native integrations for OpenAI Agents SDK, LangGraph, CrewAI, Google ADK, Strands and Microsoft Agent Framework, plus A2A multi-agent tracing added in May 2026, and its Annotation Queues became generally available in August 2026. Datadog shows tools used and handoffs for OpenAI Agents SDK, LangGraph and CrewAI in Python, auto-instruments Claude Agent SDK, Google ADK, Pydantic AI and Strands, and uses its Annotation Queues to validate LLM judges and build golden datasets.
Galileo can run on premises, including air-gapped, as a commercial deployment, or as SaaS in Splunk Observability Cloud. Datadog documents no self-hosted option and Agent Observability is not available on its US government sites.
As of 20 October 2026, Datadog's Free plan covers 40K LLM spans a month with 15-day retention, and Pro is $160 a month billed annually for the first 100K LLM spans, then $3.50 per 10K. Splunk lists Agent Observability from $100 a month per 1.2 million spans a year, billed annually, with 500 million Luna tokens and five Luna fine-tuning runs a year, and no free tier. Datadog counts only spans that call an LLM provider, while Splunk states its figure in spans, so convert both to your own traffic before comparing. See Datadog's pricing and Splunk's pricing.
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. Of these two, Galileo's Agent Control is the closest to that, though it blocks content rather than routing actions for approval; Datadog's blocking sits in AI Guard, a separate product. Prefactor is self-serve with a free tier and is not tied to Datadog or Splunk, but it has no prompt playground or prompt registry and is hosted. 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.