Compare AI agent builders
Pick up to 3 tools. Same rows for all, prices checked Nov 15, 2026; the best score of your selection is outlined.
| Compared |  Activepieces9.4/10From $20/mochecked Nov 15, 2026 |  Langflow9.0/10Free, open sourcechecked Nov 15, 2026 |  StackAI8.8/10Price on requestchecked Nov 15, 2026 |  Dify8.7/10From $59/mochecked Nov 15, 2026 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Type | Low-code, source-available | No-code, open source | No-code | Developer tool, open source | No-code | No-code | Low-code, source-available | No-code | No-code | Low-code | Low-code | Low-code | Low-code |
| Entry price | From $24/mo | From $20/mo | From $16/mo | Free, open source | From $29.99/mo | Price on request | From $59/mo | From $29.99/mo | From $37/mo | From $189/mo | From $200/mo | Price on request | Price on request |
| Billed in | Workflow executions | Credits (one per run) | Credits | No platform fee | Tasks (and agent activities) | Runs | Message credits | Credits, per seat | Credits | Conversations | Copilot Credits | Actions and Vendor Credits | Usage-based (rate not published) |
| Free plan or trial | Free self-hosted edition | Free plan | Free plan | Free self-hosted edition | Free plan | Free plan | Free plan | Free credits | Free trial | Free plan | Free trial | No | Free trial |
| Integrations announced | 500+ | 770+ | 3,500+ | Not stated | 9,000+ | 500+ | Not stated | 3,000+ | 400+ | Not stated | Not stated | 2,000+ | Not stated |
| AI models | Connect any model with your own key; Cloud plans also include AI models without API keys | Built-in fast, smart and frontier models, or your own AI keys on Plus and up | Make's AI Provider on every plan, or your own LLM key | All major LLMs and vector databases, with your own keys | Zapier provides the models; the AI model tier you pick changes how many tasks a step uses | LLM agnostic: StackAI says you can deploy the best-performing model for each task | Choose your models and your deployment environment | Pick the model for each task; Lindy says most major models are available | 35+ models, bring your own API keys (the orchestration fee then doubles to 16%) | AI usage is metered in dollars and included up to a monthly amount; the LLM list is on Botpress's site | Microsoft-provided models, or bring your own through Azure AI Foundry (billed separately) | Pay model costs with no markup, or bring your own API keys on every plan | Choose from the major LLM providers, or bring your own model |
| Agents or workflows | Agents and workflows | Agents and workflows | Agents and workflows | Agents and workflows | Agents and workflows | Agents and workflows | Agents and workflows | Agents | Agents | Agents | Agents and workflows | Agents | Agents and workflows |
| Knowledge base / memory | Yes | Not stated | Not stated | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| Human approval | Yes | Yes | Yes | Not stated | Not stated | Yes | Not stated | Yes | Yes | Yes | Not stated | Yes | Yes |
| Self-hosting | Yes | Yes, limited | No | Yes | No | Enterprise only | Yes | No | No | No | No | No | No |
| API / MCP | Public API, CLI, webhooks, MCP server trigger | API, SDK and MCP on every plan | Make API from Core up (60 calls a minute), MCP server on every plan | Built-in API and MCP servers for every flow | Zapier MCP, SDK, webhooks on paid plans | REST API | Every app can be published as an API | MCP support to connect any tool | Gumloop MCP and CLI | API reference and an agent development kit (ADK) | Power Platform connectors and admin center; Copilot Studio MCP in the docs | API and MCP | API, CLI and MCP |
| Built for | Technical founders and ops people who want control and a low bill at volume | Small teams that want Zapier-style flows with a free tier and an exit to self-hosting | Small teams that want the visual builder and the lowest entry price on a hosted tool | Developers who are comfortable with Python and want no license bill | Non-technical teams that need an app nobody else connects | IT teams in regulated companies; solo builders only get the free tier | Developers and technical teams building AI apps with a knowledge base | Owners and small agencies who want inbox, meetings and follow-ups handled | Go-to-market and ops teams that want agents everyone can use without paying per seat | Support teams with steady ticket volume and no wish to pay per seat | Teams already paying for Microsoft 365 with an admin who knows Azure | Sales and ops teams with budget for a guided rollout | Product and CX teams shipping a support agent to customers |
| Past the limit | Overage fees are listed as capped on Starter and Pro | Free waits for the next reset; Plus and Team pay $0.007 per extra credit | Scenarios stop when credits run out; you can buy extra credits in bundles of 1,000 or 10,000, or turn on auto-purchase | None | Past your tier you move to pay-per-task billing at a higher per-task rate, unless you turn it off (then Zaps pause) | Not stated; the next step up is a custom quote | Not stated on the pricing page | When the pool runs out, Lindy pauses credit-using actions until the next cycle; top-ups cost $10 per 1,000 credits | Agents stop at 20,000 credits; overage is off by default and bills at $0.005 a credit once you enable it | At 95% of your quota a pack of 100 conversations is added automatically, and auto-recharge cannot be turned off | On prepaid capacity, agents are blocked at 125% of the pack unless a pay-as-you-go meter covers the overage | Top-ups listed in the docs: $80 per 1,000 Actions, $20 per 10,000 Vendor Credits | Optional auto top-ups; no rate published |
| Official page | Check price at n8n | Check price at Activepieces | Check price at Make | Get Langflow free | Check price at Zapier | See plans at StackAI | Check price at Dify | Check price at Lindy | Check price at Gumloop | Check price at Botpress | Check price at Copilot Studio | See plans at Relevance AI | See plans at Voiceflow |
Ready-made comparisons: n8n vs Zapier, Make vs Zapier, n8n vs Make, Gumloop vs n8n, Flowise vs n8n. Not sure where to start? Answer 3 questions.