Botpress vs n8n

Botpress is built around resolving support tickets, not just answering questions. Its Autonomous Engine completes multi-step tasks — processing refunds, updating accounts and billing details, correcting orders — inside connected systems. Two builders split the work: Viber, a natural-language builder for standing up agents by describing them, and Studio, a visual builder with an autonomous node editor and macro menu for hard cases. Knowledge Bases ingest from docs, PDFs, Linear, GitHub, Confluence, Google Drive and Notion, with Tables for structured conversation data in 1K–500K row tiers by plan. Multi-LLM support covers Claude, GPT, Gemini, Groq and Hugging Face, and it deploys across Webchat, WhatsApp (including native Carousels and Cards), Slack, Instagram, Messenger, Discord and voice. Recent additions: AI draft replies in Desk, a Gmail channel, guided migration off legacy email, linked tickets and SLA alerts.

n8n is a general automation canvas where each step's input and output sit next to its settings. You get 500+ pre-built integration nodes, custom HTTP and GraphQL requests, cURL import, and JavaScript or Python code steps wherever the visual layer runs out of road. For AI specifically: a multi-agent builder, RAG pipeline support, traceable reasoning (prompt, response, next action per step), human-in-the-loop approvals, rule-based guardrails, structured input/output enforcement, and native AI evaluation — plus a recent blog guide on prompt testing frameworks for production workflows. MCP support lets you connect legacy systems and even build/update n8n workflows from inside your own AI app. The decisive difference: n8n's control and code options are broader, Botpress's action-taking inside support systems is deeper and more turnkey.

Both are freemium, but the metering is completely different. Botpress charges by conversations, not seats: the Team tier includes unlimited seats and unlimited agents, which flips the math for support orgs that get punished for hiring. The Free tier's 25 conversations won't test a real backlog, and high-volume low-value chat (password resets, status checks) gets expensive fast under conversation metering. Enterprise adds SOC 2, GDPR, KPMG-tested security and a HIPAA BAA, plus Tables row tiers from 1K to 500K by plan.

n8n gives you the open-source core to self-host, then paid cloud tiers — Starter, Pro, and Business — where SSO and audit logging don't start until the €667/mo Business tier. The critical ceilings are execution concurrency: 5 concurrent executions on Starter and 20 on Pro, which is a real constraint for high-volume production flows that will queue. For a simple one-to-one app handoff, n8n's Pro price is roughly five times what a basic connector would cost. So Botpress pricing scales with customer conversations, n8n pricing scales with executions and plan tier — pick the meter that matches your volume shape, and note n8n's compliance features arrive only at the top tier while Botpress markets them as enterprise-grade.