n8n only. Nothing else.
General freelance sites bury n8n under four hundred other tags. Here the entire directory is people who build production n8n systems, so a shortlist takes an afternoon rather than three weeks of filtering.
Hire n8n developers who have already shipped it.
A directory of n8n specialists who came to us directly. Verified profiles have been through a five-stage human review — identity, real workflow case studies, a technical read and a reference call. Every profile that has not says so on its face.
Review status stated on every profileFunds released on approvalAuditable contract record
Available this week
Search n8n, HubSpot, AI agents…
Abul Kaium
Marketing Head | Workflow Automation, CRM & AI Solutions | Freelancer Shubham Bhattacharya & Team
€30/hr
Open to the right project
Open to the right project
Ali El Mawla
n8n automation engineer: multi-agent WhatsApp systems, RAG on Supabase, LLM API integration
€25/hr
Available now
Available now
Aniket Mishra
Data & Insights Analyst | Power BI, SQL, Python | Market & Operational Analytics
€14/hr
Available now
Available now
Review status shown on every profileSee all
Workflows shipped across
Why a specialist marketplace
Four hundred tags is not a talent pool.
Automation work fails for boring reasons — no error handling, no ownership, no record of what was agreed. The marketplace is built around those four failure modes.
You see the work, not adjectives
Verification here means a reviewer read real workflow case studies — architecture, integrations, failure handling, measured outcome — and checked a reference. Profiles that have not been through that yet are labelled as unvetted rather than quietly padded out.
Money moves on delivery
Funds sit in a funded milestone before work starts and release when you approve the submission. Nobody is chasing an invoice and nobody is working unpaid.
A written record, not a chat log
Scope, milestones, submissions and decisions live on an auditable contract timeline. If something goes wrong, there is an actual record to resolve it against.
The review process
Five stages before a profile is verified.
Nobody buys a marketplace badge that means nothing. This is exactly what a profile has to survive to carry one. Profiles still working through it stay listed, marked as not yet vetted, so you always know which is which.
Reviewed by people who build in n8n
01
Identity and CV
A real name, a profile photo and a full CV. The CV stays private to the reviewer and never appears on the public profile.
02
Workflow evidence
At least one detailed case study: the business problem, the architecture, the integrations, how failures are handled, and what changed for the client.
03
Technical review
A reviewer who builds in n8n reads the submitted work, looking for error handling, idempotency, credential hygiene and whether the system is maintainable by someone else.
04
Reference check
We contact a previous client for at least one claimed engagement. Unverifiable outcomes are removed from the profile before it goes live.
05
Published, then monitored
Approved profiles go live and stay under review. Contract outcomes, dispute history and response times feed back into standing on the marketplace.
Curated talent
People who work in n8n every day.
Specialists who applied to us directly, with the stack and the work described in their own words. Open a profile to see whether it has been through review yet.
Marketing Head | Workflow Automation, CRM & AI Solutions | Freelancer Shubham Bhattacharya & Team
Kolkata, India
I am the Marketing Head at Freelancer Shubham Bhattacharya & Team, a freelance full stack development and technology team with 9.5+ years of overall software development experience. I have been working with the team for around 3 years, handling marketing, client communication, business development, project requirements, and solution planning. Our team builds custom web applications, mobile applications, CRM and business management systems, workflow automation solutions, API integrations, AI-powered tools, and digital platforms for businesses across different industries. We focus on understanding business processes first and then designing practical technology solutions that reduce manual work, improve productivity, automate repetitive tasks, and create better operational visibility. Our automation related work includes lead management, CRM workflows, social media automation, WhatsApp based communication flows, notifications, data processing, API integrations, follow ups, and connecting different business tools and services. As the Marketing Head, I work closely with clients to understand their business requirements, coordinate with our technical team, define the right solution, and ensure that the proposed system aligns with their business goals. We work as an independent freelance team and can support projects from requirement analysis and workflow planning through development, integration, deployment, and ongoing support.
n8nWorkflow AutomationBusiness AutomationCRM Automation
n8n automation engineer: multi-agent WhatsApp systems, RAG on Supabase, LLM API integration
Beirut, Lebanon
ICT graduate from AUST, July 2026. I build multi-agent automations in n8n. My main deployed work is a WhatsApp customer-service system for a beauty centre, running on its real line with real customers. A primary agent reads each chat and decides what to do. Three specialist agents handle bookings, contacts and questions about services. Redis keeps memory per customer and merges a burst of short messages into one reply. Whisper transcribes voice notes, and service knowledge comes from a Supabase vector store with my own chunking and embedding steps. I also work with LLM APIs, Supabase, Docker and Linux. With a classmate I co-built a live crypto exchange where AI agents trade through an MCP server we wrote.
n8nAI agentsmulti-agent workflowsLLM API integration
Data Science postgraduate (MSc, Trinity College Dublin) with over three years at Optum, part of UnitedHealth Group. My day job there was building ETL pipelines and automated reporting for digital platforms used by more than a million people a month, work that fed into $7M+ in savings. I held a 96% SLA compliance rate across 160+ tickets and worked alongside more than fifty teams across the business. I am comfortable in SQL, Python, Power BI, Excel, and Azure, and outside of work I build transformer models, reinforcement learning agents, and small automation tools. What I enjoy most is turning messy data into something people can actually decide on, and sitting between the technical and business sides of a problem.
SQLSparkPythonPandas
I build AI-powered automations, AI agents, and business workflows using n8n, LLMs, APIs, and CRM platforms. My recent contract work included 8+ production automation pipelines for speed-to-lead, CRM automation, lead handling, reporting, email workflows, and data extraction, along with AI voice agents using Retell AI. I handled the full workflow lifecycle, from process design and API integration to testing, debugging, and deployment. Most of my work has been delivered through client and contract engagements, so the workflows and systems cannot be publicly shared due to client confidentiality. I can discuss the architecture, implementation approach, and outcomes of relevant projects when needed.
n8nAI AutomationAI AgentsLLMs
AI Engineer | Generative AI | AI Automation | LLM Applications | RAG & Agentic Systems | Production AI Systems
Kathmandu
AI Engineer focused on building practical Generative AI and automation systems that solve real-world business problems. Experienced in developing and deploying LLM applications, RAG systems, AI agents, intelligent workflows, and API-driven services, with a focus on turning AI capabilities into reliable production systems. Strong hands-on experience across Python, FastAPI, LLM application development, AI automation, workflow orchestration, cloud infrastructure, Docker, APIs, vector databases, and AI evaluation. Comfortable working across the full lifecycle, from designing AI workflows and integrating models and tools to deploying and monitoring production systems. AWS Certified Solutions Architect – Associate, with a strong interest in Applied AI, Generative AI, agentic systems, and production AI engineering. Always focused on one thing: building AI systems that are useful, reliable, and actually work in production.
n8nLangChainRAGmulti‑agent systems
I build end-to-end n8n systems that turn manual, repetitive business processes into fully automated pipelines, most of them with AI woven into the decision-making core, not just moving data between apps. Some examples: a content generation and publishing system for a marketing agency, where n8n orchestrates the full pipeline from a raw idea to SEO-optimised article drafts to auto-adapted LinkedIn posts, tweets, and email newsletters, publishing or scheduling on demand. A recruitment screening pipeline for Utiva that validates and deduplicates applications, sends CVs to Claude for scoring and a hiring recommendation, and routes everything through an approval flow to the hiring manager, entirely hands-off until a human decision is needed. I've also built lead generation systems that scrape, verify, and score leads before generating personalised outreach sequences, and reporting automations that pull data from multiple sources and consolidate it into a live dashboard with a single trigger. My clients are typically small agencies, sales teams, and operations-heavy businesses with a real bottleneck they need solved, not a nice-to-have automation. What I deliver isn't just a working workflow, it's one built to handle edge cases reliably, documented clearly, and handed off in a state the team can actually maintain without me.
n8nPythonJavaScriptTypeScript
What gets built here
The work people actually hire for.
Six categories cover most of the marketplace. If your problem does not fit one of them, post it anyway — scoping questions are free.
Zapier and Make migrations
Rebuild brittle automations as maintainable n8n workflows with shared error handling and environment-aware credentials.
AI agents and RAG pipelines
Retrieval, drafting and classification workflows with confidence thresholds and human review where it actually matters.
CRM and revenue operations
Lead enrichment, routing, deduplication and pipeline hygiene wired into HubSpot, Salesforce or Pipedrive.
Data sync and ETL
Reliable movement between databases, warehouses and SaaS APIs, with idempotency and replay built in from the start.
Monitoring and incident recovery
Audit existing workflows, then add retries, dead-letter queues, alerting and weekly health reporting you can trust.
Custom nodes and internal tooling
Purpose-built nodes, internal approval flows and back-office tooling for teams that outgrew off-the-shelf automation.
How hiring works
From a problem to a funded contract.
Four steps, and the money never moves before there is something to review. Funding a milestone is what turns a conversation into a contract.
01
Describe the outcome
Post publicly, keep it invite-only, or shortlist directly from the directory. Describe the outcome you need rather than the node list — experts propose the architecture.
- Public, private or invite-only
- Structured scope template
- No contact details required upfront
02
Compare real proposals
Proposals include the proposed approach, a milestone breakdown and a timeline. Clarifying questions run through the platform until you decide to move forward.
- Milestone-level pricing
- Architecture stated upfront
- Contact guard until funding
03
Fund the first milestone
Funding is what starts the contract. Money is held against that milestone, and full messaging plus file exchange unlock for both sides at the same moment.
- Funds held, not forwarded
- Messaging unlocks on funding
- Scope locked to the milestone
04
Review, approve, release
The expert submits against the milestone. You approve, request changes, or open a dispute. Approval releases the funds and the contract moves to the next milestone.
- Submission-based delivery
- Change requests on record
- Dispute path if needed
Honest comparison
Against a general freelance platform.
General platforms are fine for a logo or a landing page. Production automation is a different purchase, and it fails in different ways.
| General freelance platform | n8nexperts | |
|---|---|---|
| Talent pool | Every skill on earth, n8n as a tag | n8n specialists only |
| Profile evidence | Self-declared skills and star ratings | Reviewed workflow case studies |
| Vetting | Automated, or none at all | Human review at five stages |
| Payment | Hourly tracking or upfront transfer | Funded milestones, released on approval |
| Scope record | Chat history | Auditable contract timeline |
| Dispute handling | Generic support ticket | Reviewer who understands the work |
Pricing
Free to hire. One fee, on delivery.
No subscriptions, no bidding credits, no charge for posting a job or contacting an expert.
For clients
Free
No fee to post, search or hire
- Unlimited job posts and invitations
- Full directory and case study access
- Milestone protection included
- Payment processing passed through at cost
For experts
10%
Platform fee on released milestones
- No fee to apply or maintain a profile
- No bidding credits, ever
- Paid on approval, not on invoice terms
- Drops to 5% after €25k earned with one client
Payment processing is passed through at cost. Fees are shown on every milestone before you confirm it.
For n8n developers
Get listed once. Stop bidding forever.
No credits, no race to the bottom, no rewriting the same proposal forty times. Clients arrive having already read your case studies, and the fee only applies when a milestone is released.
Apply as an expertOne application, reviewed by a humanDecision within five working days
Your work does the sellingCase studies, not a bidding queue
Paid on approvalFunded before you start, released when accepted
Questions
The things people ask before hiring.
Still unclear on something? Post a job — scoping questions cost nothing.
How is this different from hiring on a general freelance platform?
General platforms optimise for volume across every skill. The n8n tag there is a filter on a database of millions, and nothing in the profile tells you whether someone has run automations in production. Here the directory is n8n specialists only, and every profile states plainly whether it has been through review or not.
What exactly does verified mean on a profile?
It means a human reviewed the identity details, the CV, at least one detailed workflow case study, and contacted a previous client about a claimed engagement. It does not mean we guarantee the outcome of your project — it means the person is who they say they are and has demonstrably done the work. A profile without that badge carries a Not yet vetted notice: it was built from a direct application, and the details on it are self-reported.
How does milestone funding protect me?
You fund a milestone before work begins. Those funds are held against that milestone rather than forwarded to the expert. When the expert submits and you approve, the funds release. If you request changes, the milestone stays open. If you cannot reach agreement, either side can open a dispute and a reviewer examines the contract record.
Why can I not share my email before funding?
Proposals and clarification messages pass through a contact guard that blocks direct contact details and external links until a milestone is funded. It keeps the pre-contract stage on record, which is what makes disputes resolvable. Once funding lands, full messaging and file exchange unlock immediately.
What does it cost?
Nothing for clients — posting, searching, inviting and hiring are free, and payment processing is passed through at cost. Experts pay a 10% platform fee on released milestones, dropping to 5% after €25,000 earned with the same client. There are no bidding credits and no subscription on either side.
How long does expert approval take?
Applications are reviewed by a person rather than a script, so timing depends on the queue and on how much evidence came with the application. The most common reason a submission comes back is a case study that describes what a workflow does without explaining how it handles failure.
Can I hire someone for ongoing maintenance rather than a project?
Yes. Retainer-style contracts run as recurring milestones — you fund a period, the expert works against it, and you approve at the end of each cycle. Most monitoring and support engagements on the platform are structured this way.
What happens if the work is not delivered?
If a submission does not meet the agreed milestone, you request changes and the milestone stays funded but unreleased. If the expert stops responding or the disagreement is fundamental, you open a dispute. A reviewer reads the contract timeline, the milestone scope and the submissions, and decides on release or refund.
Get started
Describe the outcome. We will find who can build it.
Posting is free and takes about four minutes. Most jobs receive their first reviewed proposals within a day and a half.
No card required · No fee to hire · Milestone protection on every contract