Listed whether or not they pay us. Prices from each tool's own pricing page, checked Oct 16, 2026.

Datadog LLM Observability (Agent Observability)

Renamed: Agent Observability

Datadog module tracing LLM apps and agents end to end with quality, safety and cost monitoring.

Best for Teams already running Datadog APM

Used Datadog LLM Observability and liked it? Upvote it. One upvote per person. Upvotes come from visitors, not from us, and never change "Our ranking".

What it does

Datadog LLM Observability gives teams visibility into production AI applications and agentic workloads. Teams feed in telemetry from multi step workflows, model calls, and orchestration frameworks to monitor output quality, safety, and operational performance. It is built for engineering teams shipping and operating models or agents in production.

Read more: 1 features and how people use themShow less

  1. Online evaluations

    Evaluates production model and agent outputs continuously to help teams understand output quality, safety, and performance.

    Use caseAn AI engineer tracks output quality and safety across live user requests to pick the right model for each stage of an inference pipeline.

    From datadoghq.com/state-of-ai-engineering

Written from Datadog LLM Observability's own website, checked . Each feature links to the page it comes from; images are Datadog LLM Observability's own.

Every plan

PlanMonthlyBilled yearlyPerWhat it includes
FreeFreeNot listedper accountup to 40k LLM spans/mo, 15 day retention
Pro$200$160per account100k LLM spans included; extra $3.50/10k (annual), $4.20/10k (monthly); on demand $240 base, $5/10k
EnterpriseNot listedNot listedper accountcontact sales

From Datadog LLM Observability (Agent Observability)'s own pricing page, checked . Prices are in US dollars per month; "billed yearly" is the monthly price when you pay for a year up front. Some of these figures came from third party reports and could not be confirmed on the official page.

Before you pay

Export your data
Yes
What Datadog LLM Observability saysAgent Observability offers an Export API that you can use to retrieve spans for running external evaluations. Source page

Only what Datadog LLM Observability's own pages state, with the quote behind each answer. Anything they don't publish is left out rather than guessed.

Getting started with Datadog LLM Observability

  1. Obtain your Datadog API key from organization settings if you do not have a Datadog Agent running.
  2. Install the SDK package appropriate for your application, such as ddtrace for Python, dd trace for Node.js, or dd java agent.jar for Java.
  3. Run your application with the required environment variables, including DD_LLMOBS_ENABLED, DD_LLMOBS_ML_APP, and DD_API_KEY.
  4. Make requests to your application that trigger LLM calls.
  5. View and analyze traces under the Traces tab on the Agent Observability page in Datadog.

Written from Datadog LLM Observability's own help pages: docs.datadoghq.com/llm_observability/quicksta...

Datadog LLM Observability video tutorials

LLM Observability with Datadog — End‑to‑End Demo, Metrics & Eval‑Driven DevArpita D Kuntoje · 10 min
Complete Datadog 15-Minute Demo: End-to-End Observability WalkthroughChirag Kular · 16 min
Datadog LLM Observability: Monitor and secure your AI workloadsDatadog · 7 min

Independent creators on YouTube, not Datadog LLM Observability or us. The video loads from YouTube only when you press play.

What the free plan includes

Free tier up to 40k LLM spans/month, 15 day retention (per third party snapshot)

Questions people ask about Datadog LLM Observability

How much does Datadog LLM Observability cost?

Paid plans start at $200 a month (Pro). 3 plans are listed above, from Datadog LLM Observability's own pricing page.

Is Datadog LLM Observability free?

Yes, there is a free plan: Free tier up to 40k LLM spans/month, 15 day retention (per third party snapshot).

Does Datadog LLM Observability have an API?

Yes.

How is Agent Observability metered and billed?

Agent Observability is metered and billed on the number of LLM spans ingested. An LLM span represents a single request to an LLM provider, so one agent workflow can produce several LLM spans. Source

Works with

Integrations
OpenAI, LangChain, AWS Bedrock, Anthropic, Claude Agent SDK, CrewAI, Google ADK, Google GenAI, LangGraph, LiteLLM
AI models
OpenAI GPT 4, text embedding ada 002, gpt 4o mini

Where this came from

Plans, prices and limits were copied from Datadog LLM Observability (Agent Observability)'s own pages and the other sources listed, then checked .

How we collect prices and rank tools · Is something wrong? Correct this listing

Others that do the same job

All LLM observability & evals tools

Visit site Datadog LLM Observability (Agent Observability) (opens the tool's site)