AgentOps

aid: agentops
url: https://raw.githubusercontent.com/api-evangelist/agentops/refs/heads/main/apis.yml
name: AgentOps
type: Index
image: https://kinlane-productions.s3.amazonaws.com/apis-json/apis-json-logo.jpg
tags:
- AI Agents
- Observability
- Evaluation
- Tracing
- Python SDK
- Open Source
- Agent Frameworks
description: AgentOps is an observability, evaluation, and debugging platform for AI agents. Its open-source Python SDK
  (with TypeScript support for OpenAI Agents) instruments agent runs in two lines of code, capturing LLM calls, tool invocations,
  costs, latencies, and multi-agent interactions. Sessions are visualized in a hosted dashboard at app.agentops.ai with
  time-travel debugging, waterfall views, and replay. AgentOps offers native integrations with 400+ LLMs and frameworks
  including CrewAI, AutoGen / AG2, LangChain, LangGraph, LlamaIndex, OpenAI Agents, Haystack, and Camel AI.
created: '2026-05-23'
modified: '2026-05-23'
specificationVersion: '0.19'
apis:
- aid: agentops:agentops-python-sdk
  name: AgentOps Python SDK
  tags:
  - SDK
  - Python
  - Tracing
  - Instrumentation
  humanURL: https://docs.agentops.ai/
  properties:
  - url: https://docs.agentops.ai/
    type: Documentation
  - url: https://github.com/AgentOps-AI/agentops
    type: SourceCode
  - url: https://pypi.org/project/agentops/
    type: SDK
  description: The AgentOps Python SDK is the primary entry point, installable via pip install agentops and initialized
    with two lines of code. It auto-instruments supported agent frameworks and LLM providers, supports custom traces via
    the @trace decorator, and ships events to the AgentOps backend or a self-hosted deployment.
- aid: agentops:agentops-typescript-sdk
  name: AgentOps TypeScript SDK
  tags:
  - SDK
  - TypeScript
  - JavaScript
  - OpenAI Agents
  humanURL: https://docs.agentops.ai/v2/integrations/openai-agents-js
  properties:
  - url: https://docs.agentops.ai/v2/integrations/openai-agents-js
    type: Documentation
  - url: https://github.com/AgentOps-AI/agentops-node
    type: SourceCode
  description: AgentOps' TypeScript SDK provides instrumentation for the OpenAI Agents SDK in Node.js applications, surfacing
    the same traces and metrics as the Python SDK inside the AgentOps dashboard.
- aid: agentops:agentops-dashboard
  name: AgentOps Dashboard
  tags:
  - Dashboard
  - Observability
  - SaaS
  - Replay
  humanURL: https://app.agentops.ai/
  properties:
  - url: https://app.agentops.ai/
    type: ApplicationURL
  - url: https://docs.agentops.ai/v2/usage/dashboard
    type: Documentation
  description: The hosted dashboard at app.agentops.ai visualizes agent sessions with waterfall views, time-travel replay,
    LLM cost tracking, and multi-agent interaction graphs. Supports session export and team collaboration features in paid
    tiers.
common:
- type: Website
  url: https://www.agentops.ai/
- type: Documentation
  url: https://docs.agentops.ai/
- type: Blog
  url: https://www.agentops.ai/blog
- type: Pricing
  url: https://www.agentops.ai/pricing
- type: Login
  url: https://app.agentops.ai/
- type: GitHubOrganization
  url: https://github.com/AgentOps-AI
- type: GitHubRepository
  url: https://github.com/AgentOps-AI/agentops
- type: LinkedIn
  url: https://www.linkedin.com/company/agentops-ai/
- type: Discord
  url: https://discord.gg/FagdcwwXRR
- type: Courses
  url: https://agentops.ai/courses
- type: License
  url: https://github.com/AgentOps-AI/agentops/blob/main/LICENSE
- type: Features
  data:
  - name: Two-Line Instrumentation
    description: Initialize observability with agentops.init() and automatic framework instrumentation.
  - name: Session Replay
    description: Time-travel debugging with full session and event replay in the dashboard.
  - name: LLM Cost Tracking
    description: Token counting and cost tracking across foundation model providers and agents.
  - name: Multi-Agent Visualization
    description: Visualize interactions between agents in CrewAI, AutoGen, LangGraph, and custom systems.
  - name: Waterfall Traces
    description: Time-based waterfall views of all events in a session.
  - name: Custom Traces
    description: Use the @trace decorator and OTel-aligned spans to instrument custom code paths.
  - name: Self-Hosting
    description: Self-hosted deployment available on Enterprise plans.
  - name: SOC 2 / HIPAA
    description: Enterprise compliance with SOC 2 and HIPAA available on the Enterprise tier.
- type: UseCases
  data:
  - name: Agent Debugging
    description: Inspect multi-step agent runs, tool calls, and intermediate reasoning to find failures.
  - name: Cost Monitoring
    description: Track token usage and cost per agent, framework, and provider.
  - name: Agent Evaluation
    description: Evaluate agent performance across sessions and compare versions.
  - name: Production Observability
    description: Monitor production agents with dashboards, alerts, and exports.
  - name: Multi-Agent Systems
    description: Visualize and debug coordination between agents in multi-agent frameworks.
- type: Integrations
  data:
  - name: OpenAI
    description: Native instrumentation for OpenAI Chat Completions and Responses APIs.
  - name: OpenAI Agents SDK
    description: First-class support for OpenAI Agents in Python and TypeScript.
  - name: Anthropic
    description: Instrumentation for Anthropic Claude models.
  - name: CrewAI
    description: Native CrewAI integration with multi-agent visualization.
  - name: AG2 (AutoGen)
    description: Native integration with AG2, formerly AutoGen.
  - name: LangChain
    description: Instrumentation for LangChain chains and agents.
  - name: LangGraph
    description: Trace and visualize LangGraph stateful agents.
  - name: LlamaIndex
    description: Trace LlamaIndex RAG and agent applications.
  - name: Haystack
    description: Instrumentation for Haystack pipelines.
  - name: Camel AI
    description: Native integration with Camel AI multi-agent system.
  - name: Cohere
    description: Instrumentation for Cohere model calls.
  - name: LiteLLM
    description: Capture calls routed through LiteLLM across providers.
  - name: Mistral
    description: Instrumentation for Mistral models.
  - name: Google Generative AI
    description: Instrumentation for Gemini and Vertex AI.
  - name: xAI
    description: Instrumentation for xAI Grok models.
maintainers:
- FN: Kin Lane
  email: [email protected]