n8n workflow templates that survive production.
Most free n8n templates stop at the happy path: one bad input or one API blip and the run dies quietly. These five retry their outside calls, check what they are given, never make duplicates, and report real failures to a shared error workflow that pings Slack. Import, add your accounts, switch on.
Five workflows, ready to import.
| Workflow | What it does | Accounts |
|---|---|---|
| 01 AI agent with tools | A chat agent that answers from your Google Sheet, drafts Gmail emails (never sends them) and searches the web | OpenRouter, Google |
| 02 Lead capture to HubSpot | Form or webhook to a checked HubSpot contact, matched on email, with spam dropped and a Slack alert | HubSpot, Slack |
| 03 Cross-post to dev.to | Your blog post on dev.to with a canonical link back, never posted twice | dev.to |
| 04 Daily KPI digest | Every weekday morning, your sheet’s numbers in Slack against the day before and the 7-day average | Google, Slack |
| 05 Error alerts (shared) | One error workflow for all the others: the failed step, the error and a link to the run | Slack |
Each folder holds the workflow.json to import and a README: the accounts it needs and setup in five steps or fewer. The same notes sit on the canvas, so you never have to leave n8n. No credentials ship in the files; you pick your own on import.
What the free ones skip.
A workflow that works once in the editor is a demo. The ones that run for months share a few habits, and every workflow in this pack has them. They are the same habits behind the n8n workflows we run in production.
- Retries on every outside call. Three tries, a few seconds apart, so a one-off timeout never becomes an incident.
- A failure path. A run that still fails reports to the shared error workflow, which posts the workflow, the step, the error and a link to Slack.
- Checked input. Webhooks reject a bad request with a clear reason instead of failing halfway through.
- No duplicates. Contacts are matched on email and posts on their canonical URL, so sending the same thing twice is harmless.
- An agent on a leash. The AI agent drafts emails rather than sending them, stops after 8 tool steps, and says when a tool fails instead of inventing a result.
// On every step that calls an outside service: "retryOnFail": true, "maxTries": 3, "waitBetweenTries": 5000 // On every workflow (Settings > Error workflow): "errorWorkflow": "Error alerts (shared)"
And this is what a real failure looks like when it reaches you:
Lead capture to HubSpot failed at Create or update the contact > Request failed with status code 401 HTTP status: 401 Open the failed execution -> link to the run in n8n
Error alerts first, then the rest.
- Import
05-error-alerts, paste a Slack incoming-webhook URL into its Settings step, and activate it. - Import the others: Workflows, then Import from file.
- Pick your own credentials on each step that asks, and follow the note on the canvas.
- In each workflow, set Settings, then Error workflow, to “Error alerts (shared)”, and activate.
Error workflows only fire for production runs, never for a test click in the editor. The KPI digest needs the workflow’s own timezone set, or it drifts an hour when the clocks change: here is why, and the fix.
Building an agent from a blank canvas instead? Our n8n AI agent tutorial walks through every step. Still choosing a tool? See n8n vs Zapier first. And if you would rather have it built and run for you, that is what our n8n automation services are for.
Common questions.
01Are these n8n workflow templates free?
Yes. All five are MIT licensed: use them on client work, change them, ship them. We ask for an email so we can send a backup link to the download, and nothing else.
02How do I import an n8n workflow template?
In n8n, open Workflows, then Import from file, and pick the workflow.json from the folder. Then select your own credentials on each step that asks for one, follow the short setup note on the canvas, and activate. Import the error-alerts workflow first, because the others report to it.
03Do they work on n8n Cloud and self-hosted n8n?
They use only n8n’s built-in steps, with no community nodes, so there is nothing extra to install. We built them on our own self-hosted n8n in November 2026. Slack alerts use an incoming webhook, so no Slack app or OAuth setup is needed.
04What makes an n8n workflow production-ready?
It survives the bad day, not just the demo. Outside calls retry before giving up, bad input is rejected with a reason instead of failing halfway, re-running it never creates duplicates, and a run that really fails tells someone. Most free templates stop at the happy path; these cover the rest.
05Why does the cross-posting workflow not post to Medium?
n8n’s own documentation lists the Medium API as no longer supported, and Medium has stopped issuing new integration tokens, so most people could not connect it. The workflow posts to dev.to with a canonical link back to your site, which is what keeps Google crediting the original.
06What does the AI agent workflow cost to run?
It runs on OpenRouter, so one key covers the model and the web search. With the default model, DeepSeek V4 Flash, an answer costs a fraction of a cent at November 2026 prices, and a web search adds about one cent. The agent stops after 8 tool steps, so a confused run cannot loop up a bill.
- Claude Code skills packTen ready-made skills, plus the install and activation guide people actually search for.Free · 10 skills
- Claude Code hooks packEight production-grade hooks with the missing documentation: events, stdin payloads, exit codes.Free · 8 hooks
- Claude Code subagents packSix specialist subagents, with the frontmatter reference and delegation patterns explained.Free · 6 subagents
- CLAUDE.md templatesFive battle-tested CLAUDE.md files: Next.js, Python, monorepo, API service, and a lean starter.Free · 5 templates
- Claude Code statuslineA one-file statusline showing model, git branch, context left and session cost. Paste, done.Free · 1 script
Want these wired into your stack?
Templates get you to a working first version. Connecting n8n to your CRM, your inbox and your own data, with monitoring that reaches the right person, is the work we do for teams: the senior engineers who scope it are the ones who build it.