Agent Frameworks

Frameworks and platforms for building AI agents.

Tool
Category
Segment
Framework / Platform
Plan
Monthly Price USD
Pricing Model
Free Tier / Trial
Included Runs / Usage
Primary Language / Runtime
Agent / Workflow Capabilities
Tool / MCP / API Support
State / Memory / Persistence
Observability / Evals
Deployment / Hosting
Team / Governance
Best Fit
Main Limits / Caveats
No tagline
Agent FrameworksGraph-based agent frameworkLangGraphOpen source$0 softwareOpen-source framework; model and infra costs separateNo LangGraph software usage meter when self-hosted; pay model provider, database, queue and hosting costsPython and JavaScript/TypeScriptStateful graphs, multi-step agents, multi-agent workflows, human-in-the-loop, persistence and streamingTool calling through LangChain/LangGraph nodes; MCP and deployed API patterns through LangSmith platformCheckpointing and stores via local/DB-backed implementations; app owns persistence in OSSLocal tracing/evals require added tools such as LangSmith, Langfuse or custom telemetrySelf-host anywhere; deploy as your own serviceNo built-in hosted team governance in OSSTeams needing explicit control over long-running stateful agent workflowsFramework is powerful but lower-level than role-based multi-agent frameworks; production ops are yours
No tagline
Agent FrameworksManaged agent platformLangSmith / LangGraph PlatformPlus$39/seat/moSeat plus usage for traces, deployments, Fleet, Engine and SandboxesNo; paid plan10K base traces/month; 1 free dev-sized agent deployment; 500 Fleet runs/month; unlimited Fleet agentsHosted platform for Python/JS LangGraph agentsDeploy and manage long-running agents; Fleet agents; Engine and Sandboxes access30+ deployment API endpoints, MCP server exposure and remote MCP tools per pricing pageLangSmith Cloud; hybrid/self-hosted options require EnterpriseTracing, evals, monitoring, annotation queues, Engine issue diagnosis and sandbox usageCloud; production uptime billed per minute; deployment runs billed after included dev deploymentUnlimited seats; up to 3 workspaces; email supportTeams shipping LangGraph agents with hosted deployment and observabilityModel usage billed separately; production uptime/runs/sandboxes add metered costs
No tagline
Agent FrameworksManaged agent platformLangSmith / LangGraph PlatformEnterpriseCustomCustom enterprise packageNo public self-serve free tier beyond DeveloperCustom seats, workspaces, Fleet, Engine and Sandbox packagesHosted/hybrid/self-hosted LangSmith platformEnterprise deployment, observability, evals, Fleet, Engine and SandboxesLangGraph APIs, MCP server exposure, remote MCP tools and custom packagesCloud, hybrid or self-hosted; self-hosted data plane/VPC optionsEnterprise tracing/evals/monitoring and support SLACloud, hybrid or fully self-managedCustom SSO, RBAC, SLA, trainings and architectural guidanceEnterprises standardizing on LangGraph with strict data/governance needsRequires sales; costs depend on hosting model and usage
No tagline
Agent FrameworksRole-based multi-agent frameworkCrewAIOpen source$0 softwareOpen-source Python framework; model and infra costs separateNo framework usage meter when run locally/self-hosted; you pay LLM/tool/hosting costsPythonAgents, crews, flows, tasks, processes, memory, knowledge, guardrails and human-in-the-loop patternsCrewAI tools, LangChain tools, custom tools and framework adaptersMemory and knowledge features available in framework; app owns storage/runtimeOpenTelemetry, tracing, LLM testing and guardrails via platform/docs depending setupLocal/self-hosted or own cloudOpen-source project governance; team controls depend on your infraBuilders who want role/task-oriented multi-agent orchestration without hand-authoring graph state machinesAbstractions can be less explicit than graph-first orchestration for complex deterministic flows
No tagline
Agent FrameworksHosted agent workflow platformCrewAIFree$0Hosted workflow execution plan50 workflow executions/month; unlimited deployments listedCrewAI cloud / AMP-style hosted platformVisual editor, AI copilot, deployments, tracing, OpenTelemetry, LLM testing, guardrails and human inputGitHub integration; integration toolkit; Slack/Teams workflow chat requires EnterpriseHosted CrewAI cloud; enterprise/private infra only on EnterpriseTracing, performance metrics, token count, hallucination scores and deployment historyCrewAI cloudCommunity support; enterprise adds RBAC/SSOTesting CrewAI workflows and validating whether AMP fits a teamHard 50 execution cap; no dedicated support or enterprise controls
No tagline
Agent FrameworksHosted agent workflow platformCrewAIEnterpriseCustomEnterprise hosted/private infrastructure planRequest trialUnlimited maximum executions; included executions sized to workflow; flexible overageCrewAI cloud or customer/private infrastructureEverything in Free plus enterprise deployment, scale, dedicated support/training and workflow chat integrationsGitHub integration, enterprise integrations, Slack/Teams support and private infrastructure optionsCrewAI or customer infrastructure; dedicated VPC, NAT, FedRAMP High options shownTracing, OpenTelemetry, usage dashboard, LLM testing, guardrails and metricsCrewAI cloud, private infrastructure or customer environmentSSO, RBAC, dedicated Slack/Teams support, onsite/training/development supportLarge organizations operationalizing agentic workflows with support and private infrastructureNo public list price; execution sizing and overage are sales-led
No tagline
Agent FrameworksAPI-native agent SDKOpenAI Agents SDKSDK$0 SDK; API/tool usage billedOpen-source SDK plus OpenAI API token/tool pricingSDK is free; API requires paid usage unless account credits applyNo SDK run quota; usage billed through model tokens and tools such as web search, file search and containersPython and TypeScriptAgents, specialist handoffs, guardrails, human review, streaming, traces, voice agents and sandbox agentsFunction tools, hosted tools, MCP integrations and OpenAI platform toolsApplication owns state/storage; SDK supports resumable state and run resultsBuilt-in tracing for LLM generations, tool calls, handoffs and guardrailsSelf-host orchestration; controlled sandbox capabilities available via API/SDKGovernance depends on API org/project controls and app implementationApps already using OpenAI models that want a native, typed agent loop with handoffs and tracingTightly aligned with OpenAI APIs; model/tool costs can dominate, and long-lived infrastructure is app-owned
No tagline
Agent FrameworksAPI-native agent runtimeOpenAI API ToolsUsage-basedUsage-basedPer-token and per-tool API pricingNo standing free tier listed on pricing pageExamples: web search $10/1K calls; containers from $0.03 per 20-min 1GB session; file search $0.10/GB/day after 1GB free plus $2.50/1K tool callsOpenAI API runtimeModel calls plus hosted tools for agent workflowsWeb search, hosted shell/code interpreter containers, file search and other platform toolsFile storage/tool state billed separately; app controls broader persistenceUsage can be traced through Agents SDK and platform logsOpenAI-hosted tools plus your application serverAPI account/project controls; enterprise controls depend on contractBudgeting OpenAI-native agents beyond SDK-level software costPrices vary by selected model, context length and tools; verify before high-volume workloads
No tagline
Agent FrameworksOpen-source agent SDKGoogle ADKOpen source$0 softwareOpen-source framework; model/cloud costs separateNo ADK framework quota when run locally; pay Gemini/API/cloud/runtime costs separatelyPython, TypeScript, Go, Java and KotlinAgents, graph workflows, tools, memory, callbacks, plugins and multi-agent patternsBuilt-in tools, function tools, third-party tools, MCP tools and REST/API referencesMemory and persistence patterns in ADK; deployment runtime can use Google Cloud servicesEvaluation/debug/dev UI capabilities in ADK ecosystemLocal, self-hosted or Google Cloud/Vertex AI Agent EngineEnterprise controls come from Google Cloud when deployed thereTeams wanting a Google-backed, multi-language agent SDK optimized for Gemini but not limited to local prototypingBest fit if you are comfortable with Google/Gemini ecosystem; cloud deployment costs are separate
No tagline
Agent FrameworksManaged agent runtimeVertex AI Agent EngineUsage-basedUsage-based after free tierCompute, memory, sessions and memory-bank usageMonthly runtime free tierRuntime free tier: first 180,000 vCPU-seconds and 360,000 GiB-seconds per month/project; then vCPU $0.0864/hr and RAM $0.009/GiB-hrGoogle Cloud managed runtime for ADK and agent appsScale agents in production; runtime plus code execution, sessions and memory bank servicesGoogle Cloud APIs, Vertex AI SDK, tool integrations and cloud IAMSessions billed per stored event; memory stored/retrieved billed separately after free retrieval allowanceCloud monitoring/logging and Vertex AI toolingGoogle Cloud managed runtimeIAM, regions, VPC/security controls depend on Google Cloud configurationDeploying ADK/agent workloads to managed Google Cloud infra with a small runtime free tierCode execution, sessions and memory bank started billing from Jan 28, 2026; model calls billed separately
No tagline
Agent FrameworksMicrosoft agent SDKMicrosoft Agent FrameworkOpen source$0 softwareOpen-source SDK; model/cloud costs separateNo framework usage meter; pay Azure OpenAI/OpenAI/Azure AI/model and hosting costs separately.NET and PythonUnified foundation for agents and graph-based workflows, extending ideas from Semantic Kernel and AutoGenTools, Agent Skills, providers, A2A protocol and Azure Functions/Durable Task integrationsMemory and persistence patterns; app/runtime owns storageDevUI and integration with Microsoft/Azure observability patternsSelf-hosted app, Azure Functions, Durable Task or other .NET/Python hostingEnterprise governance via Azure/Microsoft stackTeams in Microsoft/.NET/Azure ecosystems building future-facing agent workflowsNewer unified framework; migration may be needed from AutoGen/Semantic Kernel projects
No tagline
Agent FrameworksLegacy/open multi-agent frameworkMicrosoft AutoGenOpen source$0 softwareOpen-source framework; model/cloud costs separateNo framework usage meter; pay model/tool/hosting costsPythonMulti-agent conversation framework with AgentChat, teams, extensions and AutoGen Studio ecosystemOpenAI/Azure/OpenAI-compatible model clients and extension APIsApplication/runtime owns persistence; newer Microsoft Agent Framework is successor directionAutoGen Studio and custom logging/tracing depending setupSelf-hosted/localOpen-source community; enterprise governance through your infrastructureExisting AutoGen projects and teams that need stable multi-agent conversationsMicrosoft now positions Agent Framework as the unified successor; consider migration for new Microsoft-stack projects
No tagline
Agent FrameworksMicrosoft LLM app frameworkSemantic KernelOpen source$0 softwareOpen-source SDK; model/cloud costs separateNo framework usage meter; pay model/tool/hosting costsC#, Python and JavaAgents, plugins, planning/orchestration, memory patterns and function/tool compositionPlugins expose app functions to LLM workflows; connectors to OpenAI, Azure OpenAI and other servicesMemory/connectors available; app owns persistenceObservability depends on Microsoft/Azure/OpenTelemetry setupSelf-hosted app or Azure/Microsoft app stackGovernance through app/Azure controlsTeams with existing .NET/Java/Python business logic to expose as agent toolsMicrosoft Agent Framework is the forward-looking unified path for agents/workflows
No tagline
Agent FrameworksTypeScript agent frameworkMastraOpen source$0 softwareApache-2.0 open-source framework; model/infra costs separateNo OSS framework usage meter; self-host and pay provider/hosting/storage costsTypeScript / JavaScriptAgents, workflows, tools, memory, instructions, durable multi-step processes and observational memoryMCP, tools, providers and deployable server/runtime conceptsMemory and storage abstractions; app owns infrastructure in OSSObservability hooks; hosted platform adds event metering and retentionHost anywhereEnterprise self-hosted features require custom licenseTypeScript teams wanting a code-first agent framework with workflows and memoryHosted platform features are separate from OSS; memory gateway/model costs are separate
No tagline
Agent FrameworksHosted agent platformMastra PlatformStarter$0/moHosted platform with included events/CPU and overages100K observability events then $10/100K; 24 CPU hours then $0.35/hr; 15-day retention; 1GB platform storageMastra hosted platformHosted agents, workflows, Studio, Server and Memory GatewayProvider/model access via Memory Gateway; BYOK supported; MCP/tools via framework1GB data storage; 250MB retrieval storage then $20/GB100K logs/metrics/traces included; 15-day retentionMastra hosted platformUnlimited users, deployments and projects; no multiple-team featuresFree hosted start for TypeScript agent projects with visible observability and CPU quotasPersistent server costs $100/project; memory/model gateway usage has separate market-rate/markup charges
No tagline
Agent FrameworksHosted agent platformMastra PlatformTeams$250/moHosted platform with included events/CPU and overagesNo; paid plan1M observability events then $8/100K; 250 CPU hours then $0.25/hr; 6-month retention; 100GB platform storageMastra hosted platformMultiple teams, SSO, SOC 2 docs, hosted agents/workflows and ServerMCP/tools/providers through Mastra framework and gateway100GB storage; 1GB retrieval storage then $20/GB1M logs/metrics/traces included; 6-month retentionMastra hosted platform; on-prem/custom via EnterpriseMultiple teams, SSO and SOC 2 docsGrowing teams standardizing on Mastra with hosted observability/runtimeCPU/events/storage overages apply; persistent server and memory gateway costs are separate
No tagline
Agent FrameworksDocument ingestion cloud for agentsLlamaIndex Cloud / LlamaParseStarter$50/moCredit-based document parsing/indexing planNo; paid plan40K credits/month; pay-as-you-go up to 400K credits; 5 usersLlamaIndex Cloud serviceDocument ingestion/parse/extract/index services for RAG/agent workflowsCloud API and LlamaIndex integrationsHosted document/indexing serviceBasic supportLlamaIndex Cloud5 usersSmall teams processing documents for RAG agentsNot a full managed agent runtime; credits can be consumed quickly by heavy parsing
No tagline
Agent FrameworksDocument ingestion cloud for agentsLlamaIndex Cloud / LlamaParsePro$500/moCredit-based document parsing/indexing planNo; paid plan400K credits/month; pay-as-you-go up to 4,000K credits; 10 users; Slack supportLlamaIndex Cloud serviceHigher-volume document ingestion for production RAG/agent workflowsCloud API and LlamaIndex integrationsHosted service; enterprise adds SaaS/hybrid deploymentSlack supportLlamaIndex Cloud10 usersTeams running significant document ingestion behind agentsFull enterprise governance/5x rate limits require Enterprise
No tagline
Agent FrameworksAI gateway for Pydantic AIPydantic AI Gateway / LogfireGatewayPricing through Logfire/accountGateway/BYOK or managed-provider inferenceSign-up based; no static public price captured in docsUsage depends on Logfire/Gateway provider billing, cost limits and chosen managed/BYOK providersPydantic AI with Logfire gatewaySingle API key across multiple providers, cost limits and native provider formatsOpenAI, Anthropic, Google Vertex, Groq, AWS Bedrock and more via gatewayGateway account manages API keys and spend limits; app owns application stateLogfire observability and enterprise features such as SSO/custom rolesHosted Logfire gatewayEnterprise features inherit from LogfireTeams wanting provider routing and spend controls around Pydantic AI agentsDocs describe gateway capabilities but do not list static plan prices on this page
No tagline
Agent FrameworksMinimal Python code-agent frameworkHugging Face smolagentsOpen source$0 softwareOpen-source Python library; model/hosting/sandbox costs separateNo framework usage meter; pay for Hugging Face Inference Providers, API models, local compute or sandbox providers as usedPythonCodeAgent, ToolCallingAgent, multi-step agents, multi-agent orchestration examples, vision/web agents and CLI toolsTools from MCP servers, LangChain, Hub Spaces and custom Python toolsAgent memory/persistence is app-owned; Hub/Spaces can host shared agents/toolsLogging/evals depend on surrounding stackLocal, Docker, Gradio Spaces, Modal, Blaxel, E2B or own infraGovernance depends on hosting/providerDevelopers wanting a small, transparent Python agent library with code-first tool executionCode execution needs sandboxing; production governance is not a full platform
No tagline
Agent FrameworksLLM program optimization frameworkDSPyOpen source$0 softwareOpen-source framework; model/infra costs separateNo framework usage meter; pay model calls during runtime and optimization/evaluationPythonSignatures, modules, ReAct/tool modules, composition, multimodal fields and optimizers for prompts/programsTools as Python functions; model/provider adapters through DSPy LM configurationProgram state is code/app-owned; no managed memory platform by defaultOptimization/evaluation metrics are core framework conceptsSelf-hosted Python app or notebook/serviceGovernance depends on app/runtimeTeams optimizing LLM pipelines and agents against metrics rather than hand-tuning promptsNot a turnkey agent hosting platform; optimization can increase model-call cost during compile/eval
No tagline
Agent FrameworksPython agent runtime/frameworkAgnoFree / OSS$0Open-source framework and local AgentOS control planeNo per-event fees for local/open-source usage; run AgentOS locally and pay model/infra costsPythonAgents, teams, workflows, memory, knowledge, storage, guardrails, human-in-the-loop and AgentOS UI100+ integrations; can serve LangGraph, DSPy and Claude Agent SDK agents through AgentOSMemory, knowledge and storage abstractions plus session historySession monitoring, metrics and system evaluations in Free pricing copyLocal AgentOS or self-hosted servicesCommunity support/forums on FreePython teams wanting framework plus local operational UI for agents/teams/workflowsLive production control plane requires Pro; exact OSS license/limits should be checked in repo for redistribution
No tagline
Agent FrameworksManaged agent control planeAgno AgentOSPro$150/moHosted/live AgentOS control plane subscriptionNo; paid plan1 live connection; 4 total seats included; unlimited monitoring, retention, knowledge, memories and chats; add seats/connections separatelyAgentOS control plane for Python agent systemsLive management of agents, teams and workflowsAgentOS exposes/runtime-manages agent systems; external framework support noted in docsUnlimited retention/knowledge/memories on pricing pageUnlimited monitoringHosted live AgentOS connection$30/mo per extra seat; $95/mo per live connectionManaging production agent systems without per-event billingEnterprise needed for self-hosted control plane, custom SSO/RBAC and mission-critical support
No tagline
Agent FrameworksFramework-agnostic TypeScript AI SDKVercel AI SDKOpen source$0 SDKOpen-source SDK; gateway/hosting/model costs separateNo SDK usage meter; pay model provider, AI Gateway, Vercel hosting/sandbox/workflow costs as usedTypeScript / JavaScriptText/structured generation, tool calling, streaming, agent patterns and UI hooks16+ providers, AI Gateway, tools, structured outputs and framework adaptersApplication owns persistence; Vercel ecosystem adds workflows/sandbox/hostingDevTools and ecosystem observability; app-level tracing/evals depend on integrationsAny JS runtime; optimized for Vercel/Next.jsGovernance depends on Vercel team/account and app codeWeb app teams building streaming AI agents and chat/product UIs in TypeScriptMore of an AI app/agent toolkit than a full multi-agent orchestration framework
No tagline
Agent FrameworksOpen-source multi-agent research frameworkCAMELOpen source$0 softwareOpen-source framework; model/infra costs separateNo framework usage meter; pay model/tool/hosting costsPythonMulti-agent society patterns, role-playing agents and agent scaling research/practiceTools/connectors through CAMEL ecosystem and custom Python integrationMemory/state depends on app and CAMEL modules usedResearch/eval tooling depends on setupSelf-hosted/localOpen-source community governanceResearch and experimental multi-agent simulations or role-based collaborationMay require more engineering for production governance compared with managed platforms
No tagline
Agent FrameworksModelScope multi-agent frameworkAgentScopeOpen source$0 softwareOpen-source framework; model/infra costs separateNo framework usage meter; pay model/tool/hosting costsPythonLLM-empowered multi-agent applications, message passing, tools, workflows and examplesCustom tools, model services and ModelScope ecosystem integrationsApplication-owned state/memory depending modules usedLogging/evaluation depends on framework modules and app setupSelf-hosted/localOpen-source community; enterprise controls are app-ownedTeams exploring open multi-agent systems in the ModelScope ecosystemLess broadly adopted in Western SaaS stacks than LangGraph/CrewAI/OpenAI Agents SDK