You build it with n8n. We run it on Cloudflare®

n8n needs a server that stays on. This engine doesn’t, and never sees your keys.

Why us

or point your assistant to mcp.indie.money

Why us

Cheaper. Safer. Closer.

Monthly

You payindie.money $0n8n Cloud Pro €50
Runsindie.money 10,000n8n Cloud Pro 10,000
AI tokensindie.money 10,000,000n8n Cloud Pro None

Security

Your keys stayindie.money In the vaultn8n Cloud Pro In the engine
What survives a runindie.money Nothing — we are serverlessn8n Cloud Pro Everything, keys included

Operations

Where it runsindie.money 335+ cities worldwiden8n Cloud Pro EU only, no choice
Scalabilityindie.money No tiers, no per-plan capn8n Cloud Pro 20 at once
Observabilityindie.money Traces and logsn8n Cloud Pro Not available
Data storageindie.money 500 MB per workflown8n Cloud Pro 200 MiB, whole account
Compare

Compare

Triage a client’s support inbox.

Every ten minutes, AI drafts each reply from their own docs. Six checks an hour is 4,320 runs a month.

n8n Cloud
€600 — a year on Pro. Starter’s 2,500 runs don’t cover it
Your own VPS
<$100 — a year, plus your time and your own AI bill
indie.money
$0 — runs and AI tokens, refilled daily
Pricing

Pricing

Most workflows run on the free tier.

Build on it. The tier refills every day, with no expiry and no promo clock.

Runs
10,000 a monththen $1 per 10,000
AI tokens · one free pot, all models
Included · 1×10,000,000 a monththen $0.15 per millionGLM 4.7 Flash and Gemma 4 26B
Frontier · 10×1,000,000 a monththen $1.50 per millionGLM 5.3, Kimi K2.7 and DeepSeek V4.1 Flash
Who runs it

Who runs it

Your assistant can publish it, price it, and cash you out.

Nothing here needs reading first. Point Claude, Cursor, VS Code or any MCP client at the endpoint — it signs itself in and can do the whole job: import a workflow, run it, price it, publish it, cash out.

One click instead: add to Cursor · add to VS Code

What runs here, and what doesn’t.

What runs today, and the gaps you’d hit in week one anyway.

Good fit

  • Scheduled, webhook, chat or AI-driven
  • Many integrations, and secrets you’d rather not host yourself
  • Small AI steps with no key to manage
  • Runs that finish within five minutes

Not a fit today

  • Postgres, MySQL, Redis, Mongo and SMTP: their protocols can’t be vault-proxied. Use their HTTP APIs instead: Supabase, Airtable, Notion, SendGrid, Gmail
  • PDF and spreadsheet extraction, file output: those formats need libraries the sandbox doesn’t carry, and files stay in memory
  • Heavy computing at the minimum run price: the computing a run may do rises with its price, so a data-crunching workflow needs more than the floor
Read the limitations