n8n Self Hosted Setup: A 7-Day Done-For-You Build
A practical guide to n8n self hosted setup for SMBs replacing Zapier: what it costs, how it's delivered in 7 days, and what to check first.

n8n Self Hosted Setup: A 7-Day Done-For-You Build
It's the second week of the month and your ops lead is manually copying leads from a HubSpot form into Pipedrive, then pinging sales on Slack because the Zapier plan capped out at 750 tasks and upgrading means another $70 a month on top of the $50 you're already paying. Multiply that across three workflows — lead intake, sales notes, invoice sync — and a 12-person agency in Austin or Manchester is quietly spending 6-8 hours a week on plumbing that should run itself.
The direct cost is the subscription creep: Zapier and Make bill per task, and task counts rise faster than headcount. The indirect cost is worse — a lead sits unrouted for two hours because someone was in a client call, and a $1,900 monthly retainer nearly opens without a reply.
This guide walks through what a proper n8n self hosted setup looks like when it's built for a real SMB, not a hobby project: the workflows, the cost, the 7-day timeline, and what you need ready before kickoff. See the n8n Automation Pack offer page for exact scope.
What n8n Automation Pack actually does
It's a hosted n8n instance — either on your own VPS or a managed cloud instance we provision — running three production workflows built around how your business already operates. Common picks: lead intake to CRM (form or ad lead → HubSpot/GHL/Pipedrive, deduped and assigned), sales notes to tasks (call notes or transcript → task in your PM tool), and invoice to accounting (a signed deal or paid invoice → QuickBooks or Xero entry).
Every workflow ships with retries on failed steps, error routing to a dead-letter path, and a Slack or Telegram alert when something breaks — because a silent failure at 2am is worse than no automation at all. A live KPI dashboard in Looker Studio or Metabase sits on top, fed directly by workflow run data, so you can see lead volume, sync failures, and processing time without opening n8n itself. You get full credential handover and a written runbook, so this is your infrastructure from day one, not ours.
How it works, step by step
- Discovery call (1 hour). We map your current tools — CRM, accounting, calendar, forms — and the exact three processes worth automating first.
- Written scope. You get a one-page document naming the three workflows, trigger conditions, and destination systems, signed off before any build starts.
- Instance provisioning. We stand up n8n on your VPS (or a managed cloud instance), with SSL, encrypted credential storage, and daily backups.
- Workflow build. Each of the three workflows is built, including error handling and alert routing, then tested against sample data from your systems.
- Dashboard wiring. The KPI dashboard connects to workflow execution logs so you can see volume and failure rate without touching n8n.
- Handover. You get admin access, a runbook, and short Loom walkthroughs of each workflow — no dependency on us to understand your own system.
- Daily operation. Workflows run on their triggers; alerts fire on failure; you request one small change per month under the care plan.
Trigger: New form submission on HubSpot — "Jordan Ellis, Ellis & Co, Interested in Q4 rollout" Action: Deduped against existing contacts → no match found → created in Pipedrive as new deal, stage "New Lead" → Slack alert to #sales: "New lead: Jordan Ellis (Ellis & Co) assigned to Maya" Action: Dashboard counter for "Leads routed today" increments; average routing time logged at 4 seconds
Before and after
| Situation | Before | After |
|---|---|---|
| New lead arrives | Manually copied from form to CRM, 1-3 hour delay | Routed to CRM in seconds, sales alerted immediately |
| Zapier task limit | Hits cap mid-month, upgrade or workflows silently stop | No task billing; n8n runs on your own instance |
| A sync fails overnight | Nobody notices until a customer complains | Slack/Telegram alert fires, error routed for review |
| Reporting on automation health | No visibility, guesswork | Live KPI dashboard shows volume and failures |
| Sales notes to tasks | Rep manually creates follow-up tasks after calls | Notes parsed and tasks created automatically |
| Changing a workflow | Requires finding original builder or rebuilding blind | Runbook and Loom walkthroughs, plus 1 monthly change included |
What it costs and what is included
| Item | Price | Notes |
|---|---|---|
| Setup (discovery, instance, 3 workflows, dashboard, docs) | $1,900 one-time | Delivered within 7 days of kickoff |
| Monthly care plan | $249/month | Monitoring, node upgrades, 1 small change/month, ops report |
| VPS or managed cloud hosting | Billed by provider | Passed through directly to you, no markup |
| AI model usage (if AI Agent tier added) | Billed by provider | Usage-based, passed through directly |
| Third-party platform fees (CRM, accounting) | Billed by provider | You keep your existing subscriptions |
No lock-in: you own the instance, the credentials, and the workflows outright. If the setup replaces even one paid Zapier or Make tier plus a few hours a week of manual data entry, the $249/month care plan typically costs less than what it replaces within the first month or two — without needing to project exact savings.
Delivery timeline
Day 0 — Kickoff
Discovery call, tool access requested, written scope signed off.
Days 1-3 — Build
Instance provisioned with SSL and backups; three workflows built and tested against sample data.
Days 4-5 — Dashboard and hardening
KPI dashboard wired to workflow logs; error routing and alert channels configured and tested.
Days 6-7 — Handover and go-live
Credential handover, runbook and Loom walkthroughs delivered, workflows switched to live triggers.
What you need to have ready
- Admin access (or a way to grant it) to your CRM — HubSpot, GHL, or Pipedrive
- Admin access to your accounting tool — QuickBooks or Xero
- A VPS you control, or approval for us to provision a managed cloud instance
- Slack or Telegram workspace access for alert routing
- Sample data: a few real (or anonymized) leads, invoices, or call notes to test against
- Sign-off from whoever owns data compliance if customer PII flows through the workflows
- A named point of contact for the discovery call and for approving the written scope
Doing it yourself vs done-for-you
| Factor | DIY (Zapier/Make/self-managed n8n) | Done-for-you (this pack) |
|---|---|---|
| Time to first working workflow | Days to weeks, depending on in-house n8n familiarity | Live within 7 days |
| Ongoing task/subscription cost | Rises with volume (Zapier/Make tiers) | Fixed $249/month, no per-task billing |
| Error handling | Often skipped until something breaks in production | Retries, dead-letter routing, alerts built in from day one |
| Hosting and backups | Your team configures and maintains SSL, backups, upgrades | Included in setup and care plan |
| Documentation | Frequently exists only in one person's head | Runbook and Loom walkthroughs handed over |
| Best fit | Teams with existing n8n/DevOps capacity and time to spare | Teams that want it running without building in-house expertise first |
Compliance and risk notes
- If customer PII (names, emails, deal values) flows through the workflows, confirm your CRM and accounting platform contracts already cover that data handling — this setup doesn't change your existing data processing agreements.
- Data residency: note where your VPS or managed cloud instance is physically hosted if your customers or regulators care about that (for example, UK/EU data staying in-region).
- Platform policies: CRM and accounting platforms may rate-limit or flag high-frequency API access; workflows are built within documented limits, but sudden volume spikes should be flagged to us.
- Credential storage: all API keys and OAuth tokens sit in n8n's encrypted credential vault, not in plain text in workflow definitions.
- This is not legal or compliance advice — if you operate in a regulated industry (finance, healthcare), have your compliance lead review the data flow before go-live.
Mistakes buyers make
- Picking workflows before mapping the actual bottleneck. Fix: let the discovery call surface the highest-cost manual process first, even if it's not the one you assumed.
- Skipping the sample data step. Fix: provide real (or realistic anonymized) test data upfront, or testing surfaces edge cases only after go-live.
- Treating the care plan's "1 small change" as unlimited. Fix: batch minor tweaks into a monthly list rather than requesting them piecemeal.
- Not naming a single point of contact. Fix: designate one person to approve scope and test outputs, to avoid conflicting feedback mid-build.
- Assuming self-hosting means zero ongoing cost. Fix: budget for VPS hosting and any model usage separately — these are passed through, not included in the flat fee.
FAQ
Is this n8n's cloud product or truly self-hosted?
It's a self-hosted instance — on your own VPS or a managed cloud server we provision for you — not n8n's official cloud plan. You get full admin access and own the instance outright.
Can I choose different workflows than lead intake, sales notes, and invoices?
Yes. Those three are the most common starting point, but the discovery call maps your actual bottlenecks and the written scope reflects whichever three processes matter most to you.
What happens if I want to cancel the care plan later?
You keep the instance, workflows, and all credentials — there's no lock-in. You'd take over monitoring and updates yourself, or pause and resume the care plan later.
Do you support HubSpot, GHL, and Pipedrive equally well?
Yes, all three are common CRM integrations for this pack, along with QuickBooks and Xero on the accounting side. Other tools can be scoped during discovery.
What's the AI Agent tier, and do I need it?
It adds one tool-using agent workflow — for example, an agent that reads an inbound email and decides which CRM action to take — on top of the three standard workflows. Most SMBs start without it and add it later once the core three are stable.
How fast is "2-14 days" versus the "7 days" mentioned here?
This specific pack is scoped to go live within 7 days of kickoff; other Bharat AI Saathi systems vary in the 2-14 day range depending on complexity.
Next step
Email support@bharataisaathi.com or message us on WhatsApp with your CRM, accounting tool, and the one manual process costing you the most time each week. Review the full scope on the n8n Automation Pack page, and see how we work before kickoff.
n8n Automation Pack
Third-party platform and model usage billed to you directly. No lock-in. Live in 7 days.