Key takeaways
- Per-seat pricing punishes adoption: The more useful an internal tool becomes, the more it costs to let people across the organization use it.
- Viewer and occasional-user seats add up: Most enterprise users consume information rather than build applications, making seat-based pricing expensive at scale.
- Hidden costs matter: SSO upgrades, audit-log tiers, seat audits, integration work, and migration costs often exceed the advertised subscription price.
- Self-hosted TCO scales with usage: Infrastructure, GPU resources for local AI, storage, and operations become the primary costs instead of employee headcount.
- Run the numbers for your organization: The cost crossover often favors self-hosting sooner than expected once all users and long-term licensing costs are included.
How per-user pricing scales painfully
The viewer-seat problem
Internal tools have a lopsided usage pyramid: a handful of builders, a modest group of regular editors, and a long tail of people who look at a dashboard, approve a request, or check a record once a week. Per-seat models bill the entire pyramid. Some vendors offer cheaper “viewer” or “end-user” tiers, but the discount rarely matches the usage asymmetry — and the moment a viewer needs to edit one field, they become a full seat.
The result is a perverse incentive that every IT leader recognizes: teams share logins, export data to spreadsheets to avoid seats, or restrict genuinely useful tools to a subset of the people who need them. The pricing model actively degrades the value of the software it prices.
Seat audits and true-up anxiety
Per-seat contracts create ongoing administrative work: provisioning and deprovisioning against the license count, quarterly reconciliations, true-up negotiations at renewal, and the awkward internal conversation about who “deserves” a seat. None of that work produces value; all of it consumes the time of the people who were supposed to be saving time with low-code.
Growth compounding
Per-seat costs compound along two axes at once: your organization grows, and your tool portfolio grows. Ten apps used by 40% of a growing company is a license line that rises every quarter with no corresponding decision point — nobody ever approves the increase; it just happens at renewal.
The hidden costs beyond the license
- Feature-gated tiers. SSO, granular permissions, audit logs, and environment management — the features enterprises need most — commonly live in the highest tiers. The effective per-seat price for an enterprise-ready deployment is often well above the advertised entry price.
- Integration and rebuild work. Apps built in a proprietary builder embody real engineering effort in a format only that vendor can execute. Every integration, query, and workflow is an investment locked to the platform.
- Vendor lock-in and pricing power. Once dozens of business processes run on a platform, the vendor holds the leverage at renewal. Migration cost is your ceiling for price increases, and vendors know it.
- Exit costs. The least visible line: if you leave, most proprietary low-code apps cannot be exported in an executable form. Leaving means rebuilding, which means the exit cost of year five was silently committed in year one.
The self-hosted, unlimited-user cost structure
The alternative model inverts the shape of the cost curve. A self-hosted platform with no per-seat fees has a real, non-trivial fixed floor — and a marginal user cost of zero. For a Megapodes deployment the components are concrete:
- Compute: stateless server and worker replicas, PostgreSQL (with pgvector), Redis, and S3/MinIO storage — sized to workload, not headcount.
- GPU host for local AI: one machine running Ollama for build-time AI and runtime AI Employees; roughly 8GB VRAM is the guidance, and CPU-only operation is possible but slower. This is the line item cloud-AI platforms hide inside their subscription — here it is visible, owned, and flat.
- Operations: Docker-based deployment, backups, monitoring (OpenTelemetry supported), and upgrade windows — real staff time, honestly counted.
- License: a flat platform license with no per-seat metering — unlimited users and apps. Enterprise support, custom SLAs, and deployment assistance are part of the same flat agreement — see pricing and the enterprise page for how that is structured.
The defining property: onboarding your 500th or 5,000th user changes the invoice by nothing. Adoption becomes purely good news.
A worked scenario (illustrative)
Consider an organization of 500 people where internal tools reach 60% of staff: 20 builders, 80 regular editors, and 200 occasional viewers.
Per-seat model. Suppose blended per-seat pricing lands somewhere in the low-tens-of-dollars per user per month once builder seats, end-user seats, and an enterprise tier (for SSO and audit logs) are averaged — a typical shape for this category. At 300 billed users, even a conservative blended rate puts the annual license alone in the low-to-mid six figures. Add seat administration, tier upgrades, and annual increases at renewal, and the line grows with every hire — while the exit cost accrues silently underneath.
Self-hosted model. The same organization self-hosting pays for infrastructure (application tier, PostgreSQL, Redis, object storage, one GPU host — for many mid-size deployments this is comparable to a modest cloud infrastructure bill), the operational time of a platform team that likely already runs similar services, and a flat platform license that does not move with headcount. Doubling the user base from 300 to 600 moves the infrastructure line only as far as actual load does — and moves the license line not at all.
The honest conclusion is not “self-hosting is always cheaper.” At very small scale, a handful of SaaS seats beats running servers. The conclusion is structural: per-seat cost grows linearly (or worse) with adoption, while self-hosted cost grows sub-linearly with load. Somewhere between those curves is a crossover, and for organizations in the hundreds of users it typically arrives earlier than the pilot-phase invoice suggested. Model it with your own numbers.
The exit-cost asymmetry
One structural difference deserves its own section. Megapodes apps are schema-validated declarative metadata — collections, pages, workflows, ACL — exportable as versioned, signed bundles with secrets excluded, importable elsewhere with diff, dry-run, and approval. Combined with flat, non-metered licensing, this changes the renewal conversation permanently: the configuration that runs your business is a portable artifact you hold, not an asset the vendor holds. Even if you never migrate, the credible ability to do so is worth real money at every negotiation. Security and portability details are covered in the Trust Center and features.
Questions to ask before signing any low-code agreement
- What does a viewer cost, and what actions convert a viewer into a paid editor seat?
- Which tier do SSO, audit logs, and environment promotion actually require?
- What is the contractual cap on per-seat price increases at renewal?
- In what form can we export our applications if we leave — and can anything else execute that export?
- If we double adoption, what happens to the invoice?
Frequently asked questions
Why does per-seat pricing hurt internal tools more than other software?
Because successful internal tools spread. A tool built for a 10-person team gains viewers, approvers, and occasional users across the whole organization, and per-seat models bill every one of them. The better your internal tools are, the more they cost — a structure that penalizes exactly the adoption you want.
What are the main hidden costs beyond the per-seat license?
Seat administration and audits, feature-gated tiers that force upgrades for SSO or audit logs, integration work locked to a proprietary format, and exit costs when apps must be rebuilt elsewhere. Exit cost is the least visible and often the largest.
What does a self-hosted deployment actually cost to run?
Compute for the application tier and PostgreSQL, a GPU host for local AI inference (roughly 8GB VRAM guidance for Megapodes; CPU-only is possible but slower), storage and backups, and staff time to operate it. These costs scale with usage and data, not user count — adding your thousandth user costs nothing.
Does Megapodes charge per user?
No. Megapodes is a commercial self-hosted platform licensed as a flat platform license with no per-seat metering — unlimited users and apps at no additional cost. Your total cost is the flat license plus your infrastructure; enterprise support, custom SLAs, and deployment assistance are part of the flat agreement, not priced per user.
Model the TCO for your organization
Bring your user counts and current tooling. We will walk through the deployment footprint and the flat cost structure against your numbers.
Self-hosted — your data never leaves your network · Every change human-approved and audit-logged · SSO/OIDC available · We support your security review
