Where self-hosted n8n tends to cost less, and where it does not
Self-hosted n8n tends to cost less in subscription fees, but only if you run the Community edition or buy a Business or Enterprise license for a reason you can document. Tray.ai is sold through a sales quote, so no list price exists to set against n8n's published numbers. Your renewal decision depends on the figures in your own Tray.ai order form, which should state the task or feature tiers, any regional hosting add-on and whether on-premises options are included. Ask your account team to break the quote into those components before comparing anything. On the n8n side, the Community edition has no license fee. The pricing page lists Starter and Pro cloud plans at 20 euros and 50 euros per month billed annually, a Business plan at 667 euros per month billed annually that is currently self-hosted only, and Enterprise at custom pricing for self-hosted or cloud. Business is the first tier that includes SSO and LDAP, environments with Git and log streaming, so teams that need those controls for an internal audit should price Business rather than Community. Enterprise adds dedicated support with an SLA, which you should price separately if you depend on a contractual response time. Self-hosting moves cost from the vendor's subscription into your own infrastructure, implementation and staff time. Infrastructure, backups, upgrades and monitoring recur every month. Implementation and migration are largely one-time, but they consume engineering hours that someone must pay for internally or through a partner under a fixed-scope written agreement. Ask any partner for a quote that lists those phases separately and states what remains with your team after handover. The cheaper option on paper is often the one where the on-call and upgrade work lands on engineers who were not budgeted for it.
Which cost components change when you move from Tray.ai to n8n
This table compares each cost component of running Tray.ai against self-hosting n8n, with a note on whether each cost is one-time or recurring. Tray.ai pricing is set through a sales quote, so confirm each line in writing before comparing totals. The main takeaway is that n8n's license fee can be zero on Community, but staff time, infrastructure and Business or Enterprise license decisions still drive the real total.
| Cost component | Tray.ai side | n8n side (self-hosted) |
|---|---|---|
| Subscription or license (recurring) | Quoted by tasks and features under a sales agreement. Ask for the task volume, feature tiers and renewal terms in writing. | Community edition runs without a license key at no license fee. Business or Enterprise license keys are recurring costs quoted separately, and the Business plan is self-hosted only. |
| Regional hosting or infrastructure (recurring) | Regional hosting such as US, EU or APAC is offered, sometimes as an add-on. Confirm which regions and add-ons your contract includes. | You pay for your own servers or containers, plus the database and, in queue mode, Redis. The region is wherever you choose to run it. |
| Implementation (one-time) | Vendor or partner setup of projects, connectors and authentications, scoped in the sales agreement. | Installation with Docker Compose or a major cloud provider, plus workflow build, which you can do in-house or contract to an implementation partner under a written scope. |
| Migration (one-time) | Exporting Tray.ai projects as JSON is a starting inventory, but stored data, access control and authentication secrets are not carried over. Plan rebuilding those by hand. | Rebuilding workflows and re-creating credentials in n8n. Effort rises with the number of projects, custom connectors and callable workflows. |
| Support (recurring) | Support tiers are defined by the contract. Confirm response times and escalation paths in writing. | Community support is self-directed. Enterprise adds dedicated support with an SLA, according to n8n's pricing page, and a retainer from a services provider is an optional alternative. |
| Upgrades and patching (recurring) | Handled by the vendor as part of the subscription, with release timing set by the vendor. | Your team or a services provider applies n8n releases and tests workflows after each upgrade. Track the n8n release notes for breaking changes. |
| Internal skills and staff time (recurring) | Builders and administrators work inside the vendor's platform. Training is usually included in the agreement. | Staff time to operate n8n, manage encryption keys in queue mode, back up PostgreSQL and monitor workers. This is often the largest recurring cost. |
| On-premises connectivity (one-time setup, recurring operation) | Reaches internal systems through an on-prem agent, which is an Enterprise-only option. Confirm which connectors are supported. | A network-local n8n reaches internal systems directly, so no agent license is needed. Costs are the infrastructure and firewall work inside your network. |
What moves the comparison most
Six drivers decide whether self-hosted n8n costs less than Tray.ai over three to five years. Most of them shift cost between a recurring vendor line and the internal effort to run the platform, so check each one against your own estimates before you request a quote.
Workflow and execution volume
Tray.ai pricing is quoted by tasks and features, so the more steps your flows run, the more the bill tends to grow. n8n counts one execution as one run of a whole workflow, regardless of how many steps it contains, so long multi-step flows cost less to run on n8n than step-counted models would suggest. Ask Tray.ai how it counts tasks for your exact flows, and count your own executions per month before you compare totals.
Number of users
n8n states that all plans include unlimited users, so adding builders or business owners does not add a per-seat line. Check whether Tray.ai's quote includes per-user or per-builder fees, and whether read-only or viewer roles are charged. Department-wide rollouts often change the total more through user growth than through workflow growth.
Edition needed for identity and governance
SSO and LDAP, environments with Git, and log streaming are included only in n8n's Business and Enterprise plans, not in the free Community edition. If your security policy requires SAML single sign-on, source control for workflows, or log export, the self-hosted line item moves from zero to a paid edition. Confirm which controls you need, then price that edition alongside the Tray.ai features that cover the same requirements.
Availability and queue mode
A single n8n instance is the cheapest starting point. Queue mode splits triggers and webhooks from a pool of workers that share Redis and PostgreSQL, which adds infrastructure and operating work. Multi-main high availability is limited to Enterprise self-hosted deployments. Define your uptime and throughput requirements first, because they decide how many components you run and which license you need.
Data residency and network reach
Tray.ai offers US, EU and APAC hosting, and an on-prem agent for internal systems. Self-hosting n8n inside your own region and network removes the agent and regional add-on lines, but you then carry the cost of hosting, patching and network access. Ask Tray.ai which hosting region and agent features are included in the quote, and list which internal systems your flows must reach.
Internal platform skills
The largest hidden cost is staff time: someone must upgrade n8n, manage the encryption key shared with workers, back up the database, and troubleshoot failed executions. Tray.ai moves that work into the vendor's service, while self-hosting moves it to your team or to a support retainer. Estimate the hours your platform or IT team can give each month, and ask any implementation partner to price that work in a fixed-scope written proposal.
How n8n edition fees change the self-hosted side
The self-hosted n8n cost is not zero once you move past the free edition, because n8n's licensing splits features across editions. n8n's main codebase is under the Sustainable Use License, which allows use for your own internal business purposes. Files with .ee in their name or directory are not covered by that license and require a valid n8n Enterprise License. Without a license key, n8n runs as the free Community edition. Adding a Business or Enterprise license key enables those editions on the same installation, so the decision is about which features you need, not about reinstalling the software.
Several features move a team up a tier. SAML single sign-on is available on the Business and Enterprise plans, and n8n publishes setup guides for Okta Workforce Identity and Azure AD. Environments with Git, and log streaming, are also limited to those plans. Multi-main setup, which runs several main instances for high availability, is restricted to Enterprise self-hosted deployments and is not available in the Community edition. If your Tray.ai contract relies on SSO, version control of projects, or audit-grade log export, map each of those to a specific n8n tier before you compare totals.
Confirm current feature placement and license terms with n8n before you budget, because the pricing page and the license file can change between releases. Ask n8n in writing which edition includes each feature you depend on, whether a Business license is self-hosted only at the time of purchase, and what the Enterprise terms cover for support and SLAs. Check the n8n pricing page for the current plan list, and have your legal team confirm that your intended internal use fits the Sustainable Use License.
How to build the comparison with your own numbers
This sequence turns Tray.ai's quoted terms and your actual usage into a side-by-side model of one-time and recurring costs against self-hosted n8n over the contract term. Each step produces a number you can defend to finance and procurement.
Export and count your Tray.ai inventory
Platform owners export every project and workflow from Tray.ai as JSON and record them in a spreadsheet with owner, connectors used, trigger type and whether each one runs on the on-prem agent. Then pull monthly task volume from your own Tray.ai usage reports and note peak months separately. Done when every project has a row and you have 12 months of task history or the longest period your reports cover, with the gaps noted.
Translate task volume into n8n executions
Tray.ai bills by tasks, and a single workflow run can consume several. n8n counts an execution as one run of a whole workflow regardless of how many steps it contains, so the comparison must use executions. For each workflow, count how many runs occur per month from trigger logs or Tray.ai run history, then sum them. Done when you have a monthly execution count per workflow and a total, with the step-to-execution assumption written down for review.
Request a written Tray.ai quote with hosting terms
Ask Tray.ai's sales team for a written quote that states the task allowance, overage rates, the feature tiers required for your workflows, and the cost of each regional hosting option you would need, whether US, EU or APAC. Ask separately whether an on-premises deployment is offered, which plan it requires, and what it costs in licence and support. Done when the quote lists every line item you intend to compare, with the term length and renewal uplift stated in writing.
Select and price the n8n edition
Choose the n8n edition that matches your requirements. The free Community edition has no license cost, while the Business and Enterprise editions add SSO, LDAP, Git environments and log streaming, and multi-main high availability is limited to Enterprise self-hosted. Check n8n's pricing page for the current figures, and confirm with n8n whether your intended use fits the Sustainable Use License. Done when you have one edition selected, its current list price with the billing period, and a written note on licence fit.
Estimate infrastructure for the chosen topology
Decide between a single instance and queue mode with workers, Redis and PostgreSQL, since queue mode adds components to run. Estimate compute, storage, backup, network egress and database hosting in your own cloud or data center using your internal rate cards, and include the load balancer and monitoring you already pay for. Remember the encryption key must be shared across main and worker nodes. Done when you have a monthly infrastructure figure with each component listed and a sizing assumption you can reproduce.
Add migration, implementation and support lines
Add separate lines for rebuilding workflows, recreating authentications, replacing on-prem agent connections with a network-local n8n, testing, and cutover. Request written fixed-scope quotes from implementation partners or price internal effort, and add the cost of any support agreement you choose. Mark each line as one-time or recurring. Done when every migration and support item has its own row, a named source for the number, and a flag showing whether it is one-time or ongoing.
Price staff time to operate the platform
Estimate the ongoing hours your platform, security and integration staff spend on upgrades, backups, monitoring, credential rotation, and building new workflows, then convert that time at your internal loaded rate. Keep Tray.ai's staff time on the same basis so both sides are comparable, because Tray.ai also consumes internal administration effort. Done when operating hours per month are recorded for both options with the rate used, and the assumptions are signed off by the team that will run the platform.
Run the model across the contract term
Build one table with one-time costs in the first period and recurring costs by month across the full term you are considering, including any renewal uplift in the Tray.ai quote and infrastructure growth in the n8n scenario. Calculate totals for each option and test a higher and a lower execution volume to see which costs move most. Done when the model shows total cost of ownership for both options, the sensitivity results, and a list of the figures you still need to confirm with vendors.
What running n8n asks of your team
Running n8n yourself moves the operating work from a vendor to your own staff, so the first decision is who owns the platform day to day. The skills needed are container or Kubernetes operations to deploy and restart the service, PostgreSQL and Redis administration for the database and queue that a scaled deployment depends on, secret handling for the encryption key and stored credentials, and workflow building and testing for the automations themselves. In queue mode, the main instance's encryption key must be shared with every worker and webhook processor so they can read credentials, so whoever runs the platform must treat that key as a controlled secret and know how it is rotated. Confirm with n8n's queue mode documentation which components your chosen topology needs before you size the team.
Staff time depends less on the number of workflows than on how many of them are business-critical and how many connectors need custom nodes or code. A small set of workflows that move payroll data or trigger customer notices needs monitoring, tested change control and a documented rollback path. Connectors that Tray.ai or n8n do not cover out of the box need custom nodes or code steps, and each one adds a maintenance item that someone must own through every API change. Illustrative example: a team with twenty low-risk notification flows and a team with five flows that write to a core system will need very different hours, even though both run the same software.
Your team must own upgrades. n8n publishes release notes that you must read, test against your own workflows and schedule into a maintenance window, and no one else will do this for a self-hosted install. Plan for a test instance that mirrors production so upgrades and credential changes are rehearsed first. If you want outside help, DeployOpen offers ongoing support, upgrades and continuous testing under a retainer, and scope, deliverables and fees are agreed in a written scope before work begins. Ask for the response expectations, upgrade responsibilities and what happens to on-call coverage in that written scope.
Questions buyers ask before comparing costs
Is n8n Community edition free to run inside our company?
Yes, for your own internal business purposes. n8n's Sustainable Use License allows use or modification for internal business purposes, and without a license key the software runs as the free Community edition. The license does not cover offering n8n to third parties as a managed service or reselling it, and files with .ee in their names need a valid Enterprise License. Your costs are then infrastructure, staff time and any outside services you buy. Have counsel confirm the terms against the n8n LICENSE file before you rely on them, especially if a subsidiary or a customer will use the workflows.
What does the n8n Business plan cost and why is it self-hosted only?
n8n's pricing page lists the Business plan at 667 euros per month billed annually, and it is currently offered for self-hosted deployments only. Business adds SSO and LDAP, environments with Git, and log streaming, and SAML single sign-on is documented for the Business and Enterprise plans. Enterprise is priced on request for self-hosted or cloud and adds dedicated support with an SLA. Buy Business only if you need those features, such as identity integration or source-controlled environments. Check the n8n pricing page for current terms before you budget, because plan contents can change.
How do we estimate n8n executions for our workflows?
n8n counts one execution as one run of a whole workflow, regardless of how many steps it contains. To estimate, list each workflow's triggers (schedules, webhooks, form submissions, queue messages) and multiply by how often each fires per month. Illustrative example: a workflow that fires 1,000 times a month produces 1,000 executions, whether it has 3 nodes or 30. All plans include unlimited workflows and integrations, so execution volume shapes your infrastructure, such as worker count and database growth, more than your license fee. Measure a pilot for two to four weeks before sizing production.
How long does a Tray.ai to n8n migration take?
The duration depends on the inventory, and no reliable figure applies across estates. The main drivers are the number of projects and workflows, how many connectors have no direct n8n equivalent, how many authentications must be recreated, how many on-prem agent connections need a network-local replacement, and how much testing each cutover needs. Exporting Tray.ai projects as JSON gives you the inventory to size this. Ask for a written scope that lists these items and the phases, since most engagements are fixed-scope with deliverables and timelines agreed in writing before work starts.
Can our team run self-hosted n8n without outside help?
Yes, if your platform engineers are comfortable with containers, databases and secrets handling. n8n documents Docker Compose and major cloud providers as installation paths. Scaling is more involved: in queue mode, the main instance hands executions through Redis to workers that read and write PostgreSQL, and the encryption key must be shared with every worker and webhook processor so they can read stored credentials. High availability with multiple main instances is restricted to Enterprise self-hosted deployments. If you lack capacity, DeployOpen offers ongoing support, upgrades and continuous testing as retainers. See the n8n queue mode docs before you design the architecture.
What staff skills and time does running n8n require?
The work falls into three groups: initial build, recurring operations and occasional change. Build needs skills in Linux, containers, PostgreSQL, TLS and backup design. Recurring work includes monitoring queues and workers, applying upgrades, rotating credentials, reviewing access and responding to incidents. Change work covers new connectors, new workflows and environment promotion. The recurring share grows with the number of environments, workflows and uptime commitments. Track n8n's release notes to plan upgrades, and ask any provider to state in writing which of these tasks sit with your team and which sit with them.
How do we ask DeployOpen for a migration quote?
DeployOpen does not publish a price list. A quote follows from a written scope, and most engagements are fixed-scope, with scope, deliverables and timelines agreed in writing before work starts. To get a useful quote, send your Tray.ai project inventory, the list of connectors and on-prem agent connections, hosting and data residency requirements, user counts, expected execution volumes, and any need for Business or Enterprise features. Ask for the one-time migration work and any recurring retainer as separate line items, so you can compare the first year with later years.
When the self-hosted case holds up and when it does not
Self-hosting n8n makes financial and operational sense when a team has stable internal platform skills, when data or execution must stay in a specific region or network, and when execution volumes are predictable enough to size infrastructure once and review it periodically. In that situation the cost model shifts from a per-task or per-project quote to a mix of n8n's own edition fees, servers or Kubernetes capacity, PostgreSQL and Redis, and staff time. Teams that already run PostgreSQL, Redis and container platforms for other workloads usually carry less of the setup cost. Before deciding, confirm your expected execution profile with both vendors in writing, because n8n counts a whole workflow run as one execution regardless of steps, while Tray.ai's quote is driven by its own task and feature model.
The case weakens quickly when no one owns upgrades, backups, monitoring and incident response. Self-hosting shifts those duties to your team, and n8n's free Community edition does not include SSO and LDAP, environments with Git, or log streaming, which the Business and Enterprise plans add. Multi-main high availability is restricted to Enterprise self-hosted deployments, so a team that needs it early should price the Enterprise tier before migrating. Queue mode also requires sharing the main instance's encryption key with every worker and webhook processor, so key handling needs a documented procedure from day one. A team without that capacity is usually better served by budgeting for a managed or supported setup than by running n8n alone.
Read the Tray.ai contract for services you still depend on before assuming savings. Support tiers, professional services, implementation hours and on-premises agent licensing can sit outside the per-task quote, and removing the platform does not remove the need for those activities. List every connector, custom component and scheduled job the business relies on, then confirm which ones have a native n8n node, which need an HTTP or custom code step, and which must be rebuilt by people who know both systems. A short pilot on one department's highest-volume workflows, run in parallel for a fixed period, gives you measured effort and a realistic exit date for the contract. Keep the decision open until that evidence exists.
How DeployOpen scopes a fixed-price n8n implementation
DeployOpen is a Chainscore Labs company that plans, implements, hardens, documents and supports self-hosted open-source platforms inside your own infrastructure. For a move from Tray.ai to n8n, we start with a written scope that names the workflows, connections and environments involved, so the cost comparison in this guide becomes a concrete plan with defined deliverables.
How we scope
- Inventory every Tray.ai project, workflow and on-prem agent connection the scope covers, agreed in writing before work starts.
- Confirm which n8n edition fits each workflow group, since Community, Business and Enterprise differ on SSO, Git environments and multi-main setup.
- Record the hosting location, network paths and data residency rules that the self-hosted n8n instance must meet.
- Agree what is out of scope, such as desktop automations or connectors you plan to keep on the vendor side during a transition.
How we build
- Deploy n8n in your cloud account, data center or air-gapped network, using the Docker Compose or cloud install paths your team prefers.
- Configure queue mode with PostgreSQL and Redis where worker scale is needed, and manage the shared encryption key across processes.
- Rebuild priority workflows, re-create credentials in the target instance and test each flow against representative inputs and failure cases.
- Set up SSO, backups and monitoring for the instance, and document any configuration that depends on your network layout.
How we hand over
- Deliver architecture notes, runbooks, a workflow inventory and a cost component sheet that your finance and IT teams can review.
- Provide a cutover plan by project, with the order of migration, rollback steps and the owners who sign off each stage.
- Walk your operators through upgrades, credential rotation and common troubleshooting, so internal staff can run the platform.
How we support
- Engagements are fixed-scope, so scope, deliverables and timelines are agreed in writing before any build begins.
- Ongoing support, upgrades, backups and continuous testing are available as a retainer after handover.
- Ask for a quote that separates one-time implementation work from recurring retainer, hosting and staff time, so you can compare it with the vendor's renewal.
