Ranked · Agent platforms
The 13 best AI agents platforms (2026)
The best AI agents platforms in October 2026 are OpenClaw, Vellum and n8n.
13 agents compared · 8 with a free plan · prices checked 21 September–24 October 2026
Reviewed by The Agents Index editorial team.
Ranked: 13 of the 34 agent platforms we list. The 19 not yet ranked follow the ranking. All 34 in agent platforms →
All picks at a glance
| # | Agent | Best for | Starts at |
|---|---|---|---|
| 1 | OpenClawOpen and self-hostable: single-operator first, reachable from messaging apps instead of a dashboard. | Developers and self-hosters who already run their own infrastructure and want one personal AI agent | Free |
| 2 | VellumOpen and self-hostable, with a cloud option: a personal AI assistant, not a team workflow builder | Free, then $30/month | |
| 3 | n8nOpen and self-hostable: workflow automation first, with AI-agent nodes built invs Dify | Ops and automation teams who want to build agentic workflows visually | Free, then about $27/month |
| 4 | DifyOpen and self-hostable: an LLM-app and RAG platform first, with agents built invs n8n | Developers and teams who want an open-source, self-hostable platform to build RAG apps and agents | Free, then $59/month |
| 5 | BotpressOpen and self-hostable: a conversational agent and chat-channel platform first | Engineering-led teams building a multi-channel AI support agent who want an open-source | Free, then $189/month |
Checked between 2 September 2026 and 19 October 2026. “About” figures are converted at the European Central Bank rate of 23 October 2026.
About this list
All thirteen of these call themselves an "AI agent platform," but they split on real axes, not one ranking. n8n, Dify and Botpress are all open-source and self-hostable via Docker, and each entered the space from a different starting point. n8n started as a visual workflow-automation tool with AI-agent nodes layered on top, so its scope is any automation, not agents specifically; Dify started as an LLM-app platform with a built-in RAG pipeline, so its scope is building and shipping model-backed apps and agents; Botpress started, and stays, conversational-agent-first, so its Studio, channel deployment (WhatsApp, Messenger, Slack, Teams, Telegram, voice) and bi-directional MCP support are all built around one agent that talks to people across many surfaces, not a general workflow or RAG canvas. Of the three, only Botpress is genuinely MIT-licensed (n8n is fair-code/source-available, Dify's licence carries conditions beyond pure Apache-2.0); it is also the newest and smallest by funding, a $25M Series B against n8n's and Dify's larger rounds. Lindy and Relevance AI are both closed-source and cloud-only. Lindy is the simplest on-ramp: a no-code builder for single AI "employees" that automate recurring inbox, meeting and CRM busywork. Relevance AI is the more production-grade of the two, purpose-built for orchestrating multi-agent "workforces" with a cost-optimising LLM router, evals, tracing and RBAC/SSO.
One pattern is worth weighing across n8n, Dify, Lindy and Relevance AI before you commit: n8n's Cloud tiers and Dify's Cloud tiers both publish flat monthly or annual prices. Lindy and Relevance AI both meter usage in credits instead, and that metering has a real, documented cost: independent reviews repeatedly flag Lindy's overage charges as a billing surprise, and Relevance AI's own dual "Actions plus Vendor Credits" meter is easy to misjudge on AI-heavy agents (its pricing page now foregrounds a custom Enterprise quote over self-serve numbers). None of that makes the credit-metered platforms worse, and Relevance AI in particular ships production controls the flat-rate pair does not, but it means the real cost of scaling up is a question worth modelling before you buy, not just the number on the pricing page. Botpress sits in a third pattern of its own: a flat-published base subscription. Free costs $0 (100 conversations, 3 seats, 3 AI agents), Plus $150/month billed annually (250 conversations, $25 AI usage included) and Team $750/month (1,500 conversations, $150 AI usage included), with a 20% discount on annual billing. AI usage is INCLUDED on paid tiers rather than billed separately, so the published price covers anything below the bundled allowance.
Salesforce Agentforce sits in a category of its own: a CRM-native platform bundled into an existing Salesforce investment rather than a standalone tool you adopt on its own. No other entry here puts agents this close to live sales, service and commerce data, since every agent can call Salesforce Flows, Apex and MuleSoft directly against real records. It is also genuinely model-agnostic (GPT-5.x, Claude, Gemini via its Models API), with a real developer surface (Agent API, Python SDK, CLI). The tradeoff is Salesforce-scale complexity: pricing runs on mutually-exclusive Flex Credits or Conversations meters that have already reshaped themselves multiple times since its October 2024 launch, full cross-system context requires the separately-priced Data Cloud (rebranded Data 360), and independent buyer reports describe multi-week pilots and multi-month production rollouts, a different order of commitment than n8n, Dify, Lindy or Relevance AI.
Moveworks breaks the pattern a different way again: every other platform here CAN point its agents at customers, staff, or both, depending on configuration. Moveworks points exclusively inward: IT, HR and finance requests for a company's own employees, resolved across 100+ connected systems like ServiceNow, Workday and Okta, never customer-facing support. Its compliance bar (FedRAMP, SOC 2 Type 2, a full ISO 27000-series spread, CSA STAR Level 2) is the deepest of the eight. One material fact for evaluating it: the company is no longer independent, because ServiceNow completed a $2.85B acquisition of Moveworks in December 2025.
Glean is also employee-facing, but arrives from the opposite direction: rather than starting from a specific job (tickets, workflows, CRM records) and adding agents, it starts from permission-aware search across 100+ connected systems and a Knowledge Graph, then layers Glean Assistant (chat) and Glean Agents (a no-code builder) on top of that same indexed knowledge. That foundation is backed by real scale: a $7.2B valuation as of its June 2025 Series F, over $765M raised, and customers like Zillow (3,400+ internal agents built) and Ericsson (2,700+). It also ships MCP support in both directions today, a capability Moveworks' own blog still calls "planned." Like Moveworks, pricing is entirely sales-gated with no published rate card; independent buyer-data aggregation puts the median deal near $98,890/year.
13 agents, ranked. How we rank
Featured agents
This spot is open
Featured agents show here and at the top of their category.
- API access: Not stated
- Free tier: Not stated
- MCP server: Not stated
- Own API key: Not stated
- Open source: Not stated
- Self-hosted: Not stated
The ranking
Open and self-hostable: single-operator first, reachable from messaging apps instead of a dashboard.
Editor’s pick
Best forDevelopers and self-hosters who already run their own infrastructure and want one personal AI agent that reaches WhatsApp, Telegram, Slack, Discord, Signal and iMessage instead of a separate app.
VerdictThe most radically open and self-hosted entry in this ranking: MIT-licensed, model-agnostic, and built for a single operator rather than a team, reachable from the messaging apps you already use instead of a dashboard.
- Price
- Free · source
- Free to try
- Free plan
- Watch for
- A commenter, not the reviewer, described OpenClaw's memory as breaking down with longer-term usage, based on secondhand reports rather than direct hands-on testing
Open and self-hostable, with a cloud option: a personal AI assistant, not a team workflow builder
VerdictVellum is open source and can self-host on a user's own hardware, a VPS or a private cloud with no platform fee, or run in Vellum Cloud starting from a Free Base plan with paid packages beginning at $30/month (Mighty): model usage is pass-through, billed at cost with no markup. Unlike the workflow and multi-agent platforms elsewhere on this page, it targets one person rather than a team: a single persistent-memory assistant with a built-in skill library (email, calendar, browsing, phone calls, coding) reachable across web, macOS, iOS, Slack, Telegram and a terminal, gated by explicit Allow/Deny, risk-badged prompts before it touches files, accounts or the screen. Independent write-ups of its pricing disagree with the vendor's current plan structure, so verify current rates directly, and granting phone-calling and screen-control access to a persistent-memory assistant is a real trust surface even with permission gates in place.
- Price
- Free, then $30/month · source
- Free to try
- Free plan
- Watch for
- Some capabilities, like outbound phone calls and computer control, depend on permission approvals working correctly each time; the vendor's own site is the source for most feature claims, and independent, hands-on reviews of these specific features are not available yet.
Open and self-hostable: workflow automation first, with AI-agent nodes built in
Best forOps and automation teams who want to build agentic workflows visually while keeping data on self-hosted infrastructure.
Verdictn8n is fair-code (source-available, not OSI-approved) and free to self-host, with a visual canvas, 400+ app integrations and native AI-agent nodes so an agent can actually take actions (update a CRM, post to Slack, hit an API), not just answer a question. Cloud pricing is flat and published, from €20/month (about $22). The trade-off is scope: it is an automation platform an agent plugs into, not an agent-first product, and deeply branching agent logic can get harder to read on a canvas than in code.
- Price
- Free, then about $27/month · source
- Free to try
- Free plan
- Watch for
- Fair-code is source-available, not OSI open-source, so check the licence before commercial resale.
Open and self-hostable: an LLM-app and RAG platform first, with agents built in
Best forDevelopers and teams who want an open-source, self-hostable platform to build RAG apps and agents with full data control.
VerdictDify is open-source (Apache-2.0-derived, 149K+ GitHub stars) and self-hostable via Docker, with a built-in RAG pipeline, model-agnostic runtime across hundreds of LLMs, and a visual Workflow Studio that publishes apps as a hosted page, a REST API, an embed or an MCP tool. Cloud pricing is flat and published too, from $590/year. Its licence carries conditions beyond pure Apache-2.0 (check before commercial resale), and deeply bespoke agent logic can still be easier in a code-first framework.
- Price
- Free, then $59/month · source
- Free to try
- Free plan
- Watch for
- The "open-source" licence carries extra conditions beyond Apache-2.0, not unrestricted for every commercial use.
Open and self-hostable: a conversational agent and chat-channel platform first
Best forEngineering-led teams building a multi-channel AI support agent who want an open-source, self-hostable core with an optional managed cloud layer.
VerdictBotpress is genuinely MIT-licensed (~14.8k GitHub stars) and free to self-host, built around one conversational agent deployed across a website widget, WhatsApp, Messenger, Slack, Teams, Telegram and Twilio voice, with bi-directional MCP support (consume external MCP servers, or expose itself as one). Its managed cloud now runs Free at $0 (100 conversations, 3 seats, 3 AI agents), Plus at $150/month billed annually (250 conversations, $25 AI usage included) and Team at $750/month (1,500 conversations, $150 AI usage included), with AI usage INCLUDED on paid tiers rather than billed separately, and a 20% discount on annual billing. It is the newest and smallest by funding here ($25M Series B vs. n8n's and Dify's larger rounds) and is explicit that it is not a no-code tool: independent reviews describe needing at least one developer to wire up tool calls.
- Price
- Free, then $189/month · source
- Free to try
- Free plan
- Watch for
- Not a zero-engineering tool. Independent reviews describe needing at least one developer to wire up tool calls and integrations.
Ranked 6 to 13
| # | Agent | Price | Actions |
|---|---|---|---|
| Open and self-hostable: an all-in-one AI workspace first | |||
| 6 |  SimVerified ownerDev teams wanting an open-source, API-first workspace to build and deploy AI agents Free, then $25/seat/month | Free, then $25/seat/month | |
| Closed and cloud-only: the simplest on-ramp to a single AI "employee" | |||
| 7 |  LindyTeams that want an AI teammate embedded in Slack and a no-code app builder under the same plan From $49.99/seat/month | From $49.99/seat/month | |
| Closed and cloud-only: a shared workspace you staff, not a builder you configure | |||
| 8 |  VecbaseVerified ownerSmall teams or solo operators who want pre-built AI agent roles, shared files From $19.90/year | From $19.90/year | |
| Closed and cloud-only: breadth-first, one platform instead of many tools | |||
| 9 |  MindStudioNo-code platform for building, running and deploying AI agents as drag-and-drop workflows Free, then $20/month | Free, then $20/month | |
| Closed and cloud-only: the most production-grade, built for multi-agent "workforces" | |||
| 10 |  | Quote only | |
| CRM-native: the enterprise outlier, bundled into an existing Salesforce investment | |||
| 11 |  Salesforce AgentforceSalesforce service, sales, field-service and employee-support teams Free, then $5/seat/month | Free, then $5/seat/month | |
| Employee-facing, ticket-resolution first: never touches customer support | |||
| 12 |  | Quote only | |
| Employee-facing, search-first: a Knowledge Graph with agents layered on top | |||
| 13 |  | Quote only | |
| In this category, not yet ranked | |||
 LangChainEngineering teams building a bespoke, production-grade agent | |||
 ManusIndividuals and teams who want to offload whole multi-step tasks | |||
 ElevenLabs AgentsDevelopers and product teams who want ElevenLabs' voice quality | |||
 GC AIIn-house legal teams and corporate counsel who need fast, cited research across US case law | |||
 CrestaLarge enterprises running high-volume | |||
 LettaDevelopers and researchers who specifically need an agent whose memory, skills | |||
 GensparkA consultant, marketer, founder or operations lead who wants research, files, code, slides | |||
 AgnoPython developers who want an Apache 2.0 SDK with memory, knowledge, learning, guardrails | |||
 UiPath Agent BuilderLow-code canvas in UiPath Studio Web for building AI agents | |||
 Aisera Agent Composer (Agent Studio)Conversational, no-code builder for enterprise AI agents on the Aisera platform | |||
 StackAINo-code enterprise platform for building, deploying and governing AI agents across business systems | |||
 GumloopMultiplayer AI agent builder: teams build agents over their apps while IT keeps governance | |||
 PickaxeNo-code platform for building AI agents and selling them through branded portals | |||
 DustPlatform for building and running company agents on connected tools, data and 20+ models | |||
 LuMayEnterprise agentic AI platform with packaged voice, legal, compliance and workflow agents | |||
 PitchAI Agent EngineVerified ownerScheduled agents that run recurring office administration across accounting, CRM and mail systems | |||
 TaleVerified ownerOpen-source project workspace where teams assign tasks to AI coding agents running in sandboxes | |||
 NovastartVerified ownerInfinite-canvas OS where AI agents and teammates work in real apps, each with its own cursor | |||
 Blits.ai Agentic AI for the EnterpriseVerified ownerRegulated enterprises putting agents into banking, insurance and service desk workflows | |||
Side by side
Every field, side by side
Scroll sideways to see them all.
| Feature | OpenClaw | Vellum | n8n | Dify | Botpress | SimVerified owner | Lindy | VecbaseVerified owner | MindStudio | Relevance AI | Salesforce Agentforce | Moveworks | Glean | LangChain | Manus | ElevenLabs Agents | GC AI | Cresta | Letta | Genspark | Agno | UiPath Agent Builder | Aisera Agent Composer (Agent Studio) | StackAI | Gumloop | Pickaxe | Dust | LuMay | PitchAI Agent EngineVerified owner | TaleVerified owner | NovastartVerified owner | Blits.ai Agentic AI for the EnterpriseVerified owner |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Description | Open-source personal agent runtime you host yourself and message from any chat app. | Open-source personal AI assistant with long-term memory that acts across your tools, in the cloud or self-hosted. | Workflow automation platform for building and running AI agents on a visual canvas, with code where needed. | Open-source platform for building agentic workflows, AI agents and RAG knowledge pipelines. | AI agent platform with a visual studio, a code-first TypeScript ADK and a helpdesk, priced per conversation. | Open-source AI workspace for building, deploying and running agents across 1,000+ integrations. | Slack-native AI teammate that connects to company tools and does delegated work, billed per seat with pooled credits. | Annual membership for member-rate access to a multi-vendor model catalog, plus hosted AI agents. | No-code platform for building, running and deploying AI agents as drag-and-drop workflows. | Platform for building and running specialist business agents and multi-agent workforces. | Salesforce's platform for building, deploying and monitoring AI agents on top of CRM data. | Enterprise AI assistant and low-code agent platform for employee search and task automation. | Enterprise AI platform: permission-aware search, assistant and no-code agents over connected work apps. | Open source agent frameworks plus LangSmith, a hosted platform to trace, evaluate, deploy and govern agents. | General-purpose AI agent that plans and executes multi-step tasks on its own cloud computer. | Hosted platform for building voice and chat agents, with telephony, SDKs and a choice of LLM. | Legal AI platform for in-house counsel, spanning drafting, contract review, playbooks and research. | Enterprise contact-centre platform with autonomous voice and digital AI agents plus real-time human agent assist. | Stateful agent harness, cloud and SDK for agents with persistent, git-versioned memory. | All-in-one AI workspace with autonomous agents for slides, docs, sheets, code, media and scheduled automation. | Python SDK, self-hosted AgentOS runtime and hosted control plane for running agents as a service. | Low-code canvas in UiPath Studio Web for building AI agents that run on the UiPath automation platform. | Conversational, no-code builder for enterprise AI agents on the Aisera platform. | No-code enterprise platform for building, deploying and governing AI agents across business systems. | Multiplayer AI agent builder: teams build agents over their apps while IT keeps governance and spend control. | No-code platform for building AI agents and selling them through branded portals with billing built in. | Platform for building and running company agents on connected tools, data and 20+ models. | Enterprise agentic AI platform with packaged voice, legal, compliance and workflow agents; SaaS, on-prem or air-gapped. | Scheduled agents that run recurring office administration across accounting, CRM and mail systems. | Open-source project workspace where teams assign tasks to AI coding agents running in sandboxes. | Infinite-canvas OS where AI agents and teammates work in real apps, each with its own cursor. | Enterprise platform and delivery service for governed AI agent workflows. |
| Category | Agent platforms, General-purpose AI agents | Agent platforms, General-purpose AI agents | Agent platforms, AI workflow automation | Agent platforms, Agent builders | Agent platforms, Conversational agent builders | Agent platforms, Agent builders | Agent platforms, AI workflow automation | Agent platforms | Agent platforms, Agent builders, AI workflow automation | Agent platforms, Agent builders, AI workflow automation | Agent platforms | Agent platforms | Agent platforms | Agent frameworks, Agent platforms | Research & data agents, Agent platforms, General-purpose AI agents | Voice agents, Agent platforms | Research & data agents, Agent platforms, Industry agents, Legal agents | Support agents, Agent platforms | Agent frameworks, Agent builders, Agent memory, Agent platforms, Coding agents | Research & data agents, Agent platforms, General-purpose AI agents | Agent frameworks, Agent builders, Agent platforms | Agent platforms, AI workflow automation | Agent platforms, Conversational agent builders | Agent platforms, Agent builders, AI workflow automation | Agent platforms, Agent builders, AI workflow automation | Agent platforms, Agent builders | Agent platforms | Agent platforms | Agent platforms | Agent platforms, AI workflow automation | Agent platforms | Agent platforms |
| Pricing | From Free | From Free | From Free | From Free | From Free | From Free | From $49.99/seat/month | From $19.90/year | From Free | - | From Free | - | - | From Free | From $20/month | From Free | From $500/seat/month | - | From Free | From Free | From Free | From Free | - | From Free | From $37/month | From Free | From Free | From $249/month | From about $448/month | From Free | From Free | - |
| API | Yes | Yes | Yes | Yes | Yes | Yes | No | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | - | Yes | Yes | Yes |
| Tags |
Questions buyers ask
What is the best AI agent platform?
It depends on how much control you need over hosting and code, whether you're already on Salesforce, and whether you're building for customers or your own staff. For an open, self-hostable automation platform with AI-agent nodes bolted on, n8n. For an open, self-hostable platform built around LLM apps and RAG specifically, Dify. For an open (genuinely MIT-licensed), self-hostable platform built specifically around one conversational agent deployed across many chat channels, Botpress. For the simplest no-code on-ramp to a single AI "employee," Lindy. For production-grade, coordinated multi-agent "workforces" with enterprise controls, Relevance AI. For an enterprise already running on Salesforce CRM that wants agents acting directly on live sales/service/commerce data, Agentforce. For a deeply-certified assistant that resolves EMPLOYEE (not customer) IT/HR/finance requests, Moveworks. For permission-aware search across 100+ systems with agents grounded in that indexed knowledge, Glean.
Which AI agent platforms are open-source and self-hostable?
n8n (fair-code, source-available, free to self-host), Dify (Apache-2.0-derived, 149K+ GitHub stars, self-hostable via Docker) and Botpress (genuinely MIT-licensed, ~14.8k GitHub stars, free to self-host). Botpress is the only one of the three with no licence conditions beyond the standard MIT terms. Lindy, Relevance AI, Salesforce Agentforce, Moveworks and Glean are all closed-source and cloud-only, with no self-hosting option.
Why isn’t Sierra on this list if it’s tagged as an agent platform?
Sierra is a vertical customer-support agent product — an enterprise "agent OS" for resolving customer issues — not a general-purpose platform for building an arbitrary agent. It is already ranked as the lead entry of the dedicated /best-ai-support-agents page instead of appearing here too.
Which of these platforms has predictable, flat pricing?
n8n and Dify both publish flat monthly or annual Cloud tiers. Botpress publishes flat base tiers too (Free, Plus at $150 a month billed annually, Team at $750 a month) and includes a baseline of AI usage on paid tiers rather than billing for it separately, so the published price covers anything below the bundled allowance. Lindy and Relevance AI both meter usage in credits instead. Lindy pauses at its credit cap rather than billing overages, but credit use varies widely by task, so the bill is hard to predict for a busy workspace. Relevance AI's own dual "Actions plus Vendor Credits" meter is easy to misjudge on AI-heavy agents. Salesforce Agentforce, Moveworks and Glean are the most opaque of the nine: Agentforce meters via mutually exclusive Flex Credits or Conversations plus separate per-user licenses, Moveworks publishes no pricing at all, and neither does Glean (independent buyer-data aggregation puts its median annual deal near $98,890, ranging from $29,880 to $208,897).
Do any of these AI agent platforms have a free plan?
n8n (free Community self-hosted edition), Dify (free Community self-hosted edition, plus a free Sandbox on Cloud), Botpress (free self-hosted MIT core, plus a free pay-as-you-go tier on its managed cloud), Relevance AI (a free self-serve tier with a monthly Actions allowance) and Salesforce Agentforce (a free Foundations tier with 200,000 Flex Credits) all have a genuine free option. Lindy, Moveworks and Glean do not — Lindy offers only a 7-day trial, and neither Moveworks nor Glean has a free tier or trial at any size.
Which of these is for employees rather than customers?
Moveworks and Glean. Every other platform here (n8n, Dify, Botpress, Lindy, Relevance AI, Salesforce Agentforce) can point its agents at customers, staff, or both depending on configuration — Moveworks and Glean are both built exclusively for a company's own employees and never touch customer-facing support, though they take different approaches: Moveworks resolves specific IT/HR/ finance tickets, while Glean starts from permission-aware search and knowledge grounding across 100+ systems. Moveworks is also no longer an independent company: ServiceNow completed a $2.85B acquisition of it in December 2025.
Is there an AI agent platform built natively into a CRM?
Yes — Salesforce Agentforce. Unlike the other platforms here, which are adopted as standalone tools, Agentforce runs against a company's existing Salesforce data and requires that CRM investment underneath it. In exchange, its agents can take action directly on live sales, service and commerce records through Salesforce Flows, Apex and MuleSoft — deeper native CRM integration than any general-purpose platform in this category offers.
Sources and method
Every price here was read on the maker's own pricing page.
- OpenClaw: openclaw.ai, checked 21 September 2026
- Vellum: vellum.ai/pricing, checked 24 October 2026
- n8n: n8n.io/pricing, checked 24 October 2026
- Dify: dify.ai/pricing, checked 24 October 2026
- Botpress: botpress.com/pricing, checked 24 October 2026
All 13 sources
- Sim: sim.ai/pricing, checked 24 October 2026
- Lindy: lindy.ai/pricing, checked 24 October 2026
- Vecbase: vecbase.com/pricing, checked 24 October 2026
- MindStudio: mindstudio.ai/pricing, checked 24 October 2026
- Relevance AI: relevanceai.com/pricing, checked 24 October 2026
- Salesforce Agentforce: salesforce.com/agentforce/pricing, checked 24 October 2026
- Moveworks: moveworks.com, checked 24 October 2026
- Glean: docs.glean.com/glean-core-suite-pricing, checked 19 October 2026
The digestFree
Which agents actually ship.
New agents and articles, by email, at most once a week.
Last issue · No. 1 · 13 October 2026Devin vs Claude Code: Which Should You Pick?
One-click unsubscribe