Download$19
n8n Agent Workflows
6 import-ready n8n workflows: x402 monitor, change alerts, Stripe log, AI digest and more.
View n8n Agent Workflowsn8n AI Agent Workflows Pack
AI agents that stay on a short leash: the model extracts and writes, your rules decide.
What it is
Four importable n8n AI workflows that keep the AI on a short leash: the model extracts, classifies or writes, and the decisions stay in readable rules you can test.
| File | Trigger | What it does |
|---|---|---|
| 01-support-chat-agent.json | n8n hosted chat | AI Agent + memory + a search_faq Code Tool. Answers only from your FAQ, escalates to a human email when nothing matches |
| 02-gmail-ai-triage.json | Gmail, every 5 min | Cleans each new email (quoted replies, signatures, automated senders), classifies it (Sales lead, Support, Billing, Newsletter/promo, Other) and adds a Gmail label; sales emails also go to Slack. Never sends, archives or deletes |
| 03-ai-lead-qualification.json | Webhook (form) | Information Extractor pulls company size, budget, timeline, authority and use case from the message; a rule-based Code node scores 0-100 (hot/warm/cold with reasons); hot leads go to Slack, every lead to Google Sheets |
| 04-ai-feedback-router.json | n8n form | Customer feedback form; drops empty and link-spam posts before any AI call, Sentiment Analysis routes Positive to a testimonial-candidates sheet, Neutral to a log, Negative to rule-based urgency + a Slack alert |
Built for n8n 2.x with the LangChain nodes (versions match n8n 2.46.7), default model gpt-5-mini. Every workflow passes n8nlint --strict (checked against the real n8n node definitions, including AI connection rules) and loads in n8n's own workflow model without parameter issues.
What you get
The screenshot is real output from the kit’s own commands. Tested from a fresh download of this exact zip: ./run_tests.sh ran 13 tests OK on Python 3.13 and 3.10.
Import
1. In n8n: Import from File, pick a file from workflows/, then follow the yellow Setup note. 2. Credentials: OpenAI on each "OpenAI Chat Model" node (or replace it with Anthropic, Gemini, Ollama or another chat model node), Gmail, Slack, Google Sheets as used. 3. Replace REPLACE_ME values: store name and support email (agent), Gmail label ids (pick labels in the 4 Label nodes), spreadsheet id. python3 -m n8nlint creds workflows/ lists them. 4. Test with Execute workflow or the chat panel, then activate.
Change the rules
| Node | Settings |
|---|---|
| search_faq (Code Tool) | FAQ entries (question, answer, extra search words), MAX_RESULTS |
| Prepare email | MAX_CHARS, SKIP_SENDERS |
| Score lead | weights W, FIT_KEYWORDS, HOT / WARM thresholds |
| Prepare feedback / Urgency rules | MIN_WORDS, MAX_LINKS, URGENT_WORDS, URGENT_MAX_RATING |
Tests
./run_tests.sh # 13 tests: lint, import hygiene, Code == code/*.js, agent and classifier wiring, and the Code logic on sample data (needs Node.js; skipped without it)
./demo.sh # lint + the FAQ tool, lead scores and the feedback router on the fixturesThe tests check the deterministic parts. They do not call OpenAI; model answers vary, so try each workflow with your own data before activating.
Use with Claude Code
This folder is a Claude Code plugin with the n8n-ai-agent-workflows skill: ask Claude to add FAQ entries, change categories or scoring, and it edits, lints and re-tests.
Honest limits
- Model calls cost money on your OpenAI (or other) account. The hosted chat is public by default: anyone with its URL can use it; turn on authentication in the Chat Trigger if needed.
- AI classification and extraction make mistakes. Triage only adds labels, and the lead score is rules on extracted facts; check both before relying on them.
- The FAQ tool is keyword search over a short list, made for tens of entries. For large help centers use a vector store.
- Testimonial candidates are not published anywhere; ask the customer before you quote them. Sentiment is a model's judgment, so the Negative branch also logs every item for review.
- Emails and form messages are sent to your AI provider. Check that this is allowed for your data.
Not affiliated with n8n GmbH, OpenAI or Google. Trademarks belong to their owners.
Format and license
Zip, 196 KB, v1.0, 28 files. You may use and modify this kit for your own work and for clients you serve, including paid client work. Not affiliated with or endorsed by n8n, OpenAI or Google.
FAQ
- Do I need OpenAI?
- It's the default; you can swap in Anthropic, Gemini, Ollama or any chat model node.
- Can the triage workflow send or delete email?
- No, it only adds labels.
- Does it work on n8n Cloud?
- Yes, n8n 2.x cloud or self-hosted.
- How do I get the files?
- Right after payment, Stripe’s confirmation page shows your download link.
- Refunds?
- Email hello@youraitools.co within 14 days if it doesn’t work and we can’t fix it.
- Can I pay with crypto?
- Yes. Pay the same price in USDC or USDT on Base, Ethereum, Arbitrum, Optimism, Polygon or Solana from the "Pay with crypto" button, with no account or KYC. We check the payment on-chain and show your download links as soon as it confirms.
Get n8n AI Agent Workflows Pack
4 importable n8n AI agent workflows: FAQ support agent, Gmail AI triage, AI lead qualification with rule-based scoring, feedback sentiment router. Tested. $19.
Save with a bundle
- n8n Shopify, HubSpot & AI Workflows Bundle: 4 products for $39 ($57 separately, you save $18).
Pairs well with
Community course$39
n8n for AI Agents (community course)
Community course: build AI agent workflows in n8n you can trust. 3 modules, 9 lessons, 63 tests in real n8n.
View n8n for AI Agents (community course)Download$39
n8n Bookkeeper Invoice Exceptions Pack
n8n invoice review queue: duplicates, totals, bank-detail changes, credit notes. Never pays or posts.
View n8n Bookkeeper Invoice Exceptions Pack