You will get : I will build n8n + Claude automations for your team
Project details
I build production n8n workflows with the Claude API doing classification, extraction, and routing inside them. Three live templates (lead intake, support triage, invoice parsing) with full JSON, sample outputs, and real test data are in my connected GitHub portfolio. Not vaporware.
WHAT YOU GET: a working automation in your environment, tested against your real data before handover. All credentials wired (Anthropic, Airtable, Gmail, Slack, etc.). 14-day post-delivery bug fixes free.
WHY CLAUDE OVER GPT-4: ~10% the per-token cost on classification tasks, longer context, more predictable JSON output. Saves your API bill from day one. Hard OpenAI requirement? I can swap.
AI DISCLOSURE: workflow architecture, integration code, and all workflow logic are my work. Claude API runs inside the deployed workflow as the language model. Deliverables hand-tested before handoff.
WHAT YOU GET: a working automation in your environment, tested against your real data before handover. All credentials wired (Anthropic, Airtable, Gmail, Slack, etc.). 14-day post-delivery bug fixes free.
WHY CLAUDE OVER GPT-4: ~10% the per-token cost on classification tasks, longer context, more predictable JSON output. Saves your API bill from day one. Hard OpenAI requirement? I can swap.
AI DISCLOSURE: workflow architecture, integration code, and all workflow logic are my work. Claude API runs inside the deployed workflow as the language model. Deliverables hand-tested before handoff.
AI Algorithms
Large Language Model, Multimodal Large Language Model, Transformer ModelAI Applications
AI-Enhanced Classification, Natural Language Generation, Natural Language Understanding, Sentiment Analysis, Text RecognitionAI Models
ChatGPT, GPT-4What's included
| Service Tiers |
Starter
$250
|
Standard
$600
|
Advanced
$900
|
|---|---|---|---|
| Delivery Time | 3 days | 5 days | 7 days |
Number of Revisions | 2 | 3 | 3 |
AI Model Integration | |||
Batch Normalization | - | - | - |
Database Integration | |||
Detailed Code Comments | - | - | - |
Image Upscaling | - | - | - |
MLOps | - | - | - |
Model Deployment | - | - | - |
Model Documentation | - | - | - |
Model Monitoring | - | - | - |
Model Testing & Optimization | - | - | - |
Model Tuning | - | - | - |
Natural Language Processing | |||
NLP Tokenization | - | - | - |
Pre-Training | - | - | - |
Prompt Engineering | |||
Setup File | |||
Source Code |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$75 - $200
Additional Revision
+$50Frequently asked questions
1 review
(1)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
AM
Aleksandrs M.
Jun 15, 2026
Make.com automation broke after I added new modules, need a quick fix
Quick, knowledgeable and easy to work with. Fixed my broken Make automation and explained what went wrong so it won't happen again. Would hire again.
About Mihails
n8n + Claude API Automation Engineer | Lead Routing for SMBs
Copenhagen, Denmark - 4:49 pm local time
If your team is copy-pasting from Gmail into a spreadsheet, replying to repetitive support emails after dinner, or losing leads because nobody flagged the urgent one, that is the work I automate.
WHAT I BUILD
Lead intake: contact form fires a webhook, Claude classifies the message, urgent leads ping Slack, every lead lands in your CRM. No more "we forgot to follow up."
Support ticket triage: inbox watched, Claude tags urgency and intent, draft reply generated, handed to a human for one-click send.
Invoice and document processing: PDF in, Claude extracts vendor, amount, due date, and line items. Structured row written to Airtable, QuickBooks, or your accounting tool of choice.
Each workflow runs in n8n with the Claude API as the language model. n8n self-hosts on a $5 VPS so your data stays under your control. Claude handles classification on real-world messy inputs without hand-tuning regex.
PORTFOLIO
Three production-ready workflow templates with full JSON, screenshots, and case writeups are listed in the Portfolio section of this profile. My GitHub is connected on the sidebar; the public repo is named n8n-demo-workshop. Free to inspect, customized per client.
WHY ME
Indie software developer. Shipped Idle AI Empire, a Godot 4.x idle game, solo from concept to public release. I know how to ship working software end-to-end, not just demo prototypes.
PRICING
Three fixed-price tiers (full detail in my Project Catalog):
Starter, $250. Single workflow, single channel, basic CRM write. 1 to 3 business days.
Pro, $600. Three workflows, multi-channel routing, CRM write-back, Slack notifications. 3 to 10 business days.
Business, $900. Full intake plus classification, routing, Slack and email alerts, analytics dashboard. 5 to 14 business days.
Hourly available at $45/hr for custom scope or ongoing maintenance.
HOW I WORK
1. 20-minute discovery call or chat (free) to scope the problem
2. Fixed-price quote with deliverables and turnaround
3. Build in n8n, test against your real data, hand off
4. 14-day post-delivery support. Bug fixes free, change requests at hourly rate.
AI DISCLOSURE (per Upwork's May 2024 policy)
The deliverable is an n8n workflow that calls the Anthropic Claude API as one component for natural-language classification and extraction. Workflow architecture, integration code, and all workflow logic are my work, developed using AI-assisted tooling (Claude Code).
I do not deliver hands-off AI-generated work without human authorship and testing.
Steps for completing your project
After purchasing the project, send requirements so Mihails can start the project.
Delivery time starts when Mihails receives requirements from you.
Mihails works on your project following the steps below.
Revisions may occur after the delivery date.
Discovery call
15-minute free call to confirm scope, list integrations needed, and align on credentials timeline.
Build & wire
Build the n8n workflow architecture: trigger handling, AI classification logic, routing, and destination writes. Wire credentials for required APIs (Anthropic, Airtable, Gmail, Slack, etc.). Implement multi-item batching and defensive JSON parsing.