Integrations Overview

Arthur AI supports a broad set of frameworks and tools for building, observing, and evaluating AI applications — including OpenAI, Anthropic, LangChain, LiteLLM, CrewAI, LlamaIndex, OpenAI Agents, Google ADK, AWS Bedrock, and Mastra (JavaScript/TypeScript). To pick the right integration, identify the language you develop in (Python or JavaScript/TypeScript), the LLM framework you already use, and whether you need tracing, evaluation, or both. This page gives you a single at-a-glance reference for every supported integration so you can confirm compatibility with your stack and jump straight to the framework-specific guide.

How Arthur Integrations Work

Every Arthur integration follows the same core pattern: your application sends OpenTelemetry-compatible traces to the Arthur GenAI Engine, where they are stored, scored, and surfaced in the Arthur platform. The integration layer handles span creation, context propagation, and attribute mapping automatically — you just initialize Arthur and instrument your framework.

flowchart LR
    A[Your Application] -->|instrument| B[Arthur SDK / Integration]
    B -->|OTLP traces| C[Arthur GenAI Engine]
    C --> D[Observability Dashboard]
    C --> E[Continuous Evals]
    C --> F[Agentic Experiments]

The table below summarizes every supported integration, the language it targets, and what it provides.

Next Steps

Once you've finished following the guides above to set up tracing, you can explore these capabilities:

• Prompt Management — version and manage prompts in Arthur, then fetch them at runtime with arthur.get_prompt()
• Continuous Evaluations — set up automated quality checks that run against your traced inferences
• Agentic Experiments — run structured experiments against your AI tasks to evaluate performance at scale
• Read our Best Practices for Building Agents Blog Series — observability and tracing fundamentals for building production agents

flowchart LR
    A[Instrument Your Framework] --> B[View Traces]
    B --> C[Add Evaluations]
    B --> D[Manage Prompts]
    C --> E[Set Up Continuous Evals]
    D --> F[A/B Test Prompts]
    E --> G[Production Monitoring]
    F --> G