Major

Retool Pricing: Costs, Alternatives, and a Better Cost Model

Retool cost depends on seats, plan gates, usage, and deployment. Compare the cost model, then see how Major changes repeat-work economics.

Jason BaoUpdated
Retool Pricing: Costs, Alternatives, and a Better Cost Model

Retool pricing depends on four things: the plan you land on, the split between builders and internal users, the capabilities and deployment options gated behind higher tiers, and any usage-based items such as AI features. A starting rate tells you almost nothing about the total. Model seats, external access, AI usage and hosting together before you compare anything.

Major changes the shape of that model. Agents on Major build a deterministic app for each repeatable step, then run it as code, with hosting, database and deployment included in the platform. The expensive part, reasoning, is spent once. This guide is about cost modeling rather than a feature roundup; for that, see our Retool alternatives guide. Plan names and prices change, so everything about Retool's current plans below points to the official pricing page, checked on November 2, 2026. We print no competitor price figures here because they go stale and the page is the only source that counts.

What Retool pricing depends on

Retool prices by role. Its own pricing v2 explainer describes a model where people who build apps and people who only use them are counted differently. That explainer is dated 2023, so treat it as context for how the model works, not as current numbers.

Five cost drivers matter for a platform team.

  • Seats. How many builders, how many internal users, and whether each person's role can change over time.
  • Plan gates. Capabilities such as SSO, audit logs, or public-facing apps may sit on higher tiers. Confirm each one on the live page for the tier you would actually need.
  • External access. Customers, vendors or contractors who touch an app can be counted differently from employees.
  • AI and runtime usage. Anything metered by usage is a variable, not a fixed line.
  • Deployment and operations. Cloud versus self-hosted changes who pays for infrastructure and who carries the pager.

There is also anecdotal evidence that gates catch people off guard. A user reports on Medium that public apps were available only on higher tiers and called those tiers expensive. A commenter on Hacker News wrote, "the retool pricing is so annoying!" Both are individual reports and may be dated. Use them as a prompt to check gates, not as proof of today's policy.

Retool alternatives to compare on total cost

The comparison below is about cost model, not features. Each entry names the cost structure to investigate and the question to ask a vendor. Verify everything on the linked official page before it goes in a budget.

1. Major

Major is the enterprise platform where agents build the software they run on. A person can describe an app and ship it, and an agent can do the same for its own repeatable work. SSO, permissions, audit, database and storage are handled at the platform layer, so they are not separate line items to negotiate. The cost argument is structural: model reasoning is front-loaded when the app is built, then repeat runs execute as code. Actual spend still depends on your plan and how much judgment your workflows need. The question to ask: how much of this workflow is repeatable enough to run as code?

2. Retool

The baseline. A low-code builder for internal tools with role-based seats and tiered capabilities, per the official page. Each app needs a person to build and maintain it, so cost scales with builders and users. The question to ask: which tier holds every control we need, and how are occasional and external users counted?

3. Appsmith

An open-source-rooted internal app builder with a self-hosted option; plans are on its pricing page. Self-hosting shifts cost from licenses to your own infrastructure and operations. The question to ask: who owns upgrades, uptime and security patching?

4. Budibase

Another open-source-rooted builder with self-host and cloud options, listed on its pricing page. The same trade applies: license cost against operating cost. The question to ask: which governance features are available on the plan we would deploy?

5. ToolJet

An open-source-rooted low-code platform with plans on its pricing page. The question to ask: how are builders, end users and any AI features counted on the plan we need?

6. Superblocks

An internal app platform with plans on its pricing page. We omit specific plan claims because we could not verify them for this article. For a head-to-head on the incumbent, see Retool vs Superblocks. The question to ask: what is the unit of billing, and what is excluded from it?

The common thread in entries two through six is that a person builds each app and the app is an interface. That is the category's strength, and it is also why cost tracks headcount. If you are still choosing a category, read our guide to choosing a low-code app builder.

How to compare a Retool quote fairly

A fair comparison is the cost to operate a governed app program, not one seat rate against another. Tier gates matter as much as the monthly figure. Build a worksheet with these rows, and fill in only what the vendor has put in writing.

Hypothetical team profile, unpriced: a platform group supporting finance, support and RevOps, with a handful of builders, a few hundred occasional internal users, and some vendors who need limited access.

  • Builders. Count today and in twelve months. Note who qualifies as a builder.
  • Internal users. Split daily users from occasional ones.
  • External users. Customers, vendors, contractors. Ask how they are counted.
  • Required controls. SSO, role-based access, audit logs, environments. Record which tier holds each.
  • AI and runtime. What is metered, and what is the unit?
  • Hosting and deployment. Cloud or self-hosted, and who pays for and operates it.
  • Support. Response commitments and who owns incidents.
  • Operational ownership. Who maintains each app after the person who built it moves teams?

Mark each cell as list price, quoted, or unknown. Do not add up hypothetical rates. A sum built on unknowns looks precise and is not.

How to build a cost-aware app workflow on Major

Here is a concrete example for a finance operations team: an invoice reconciliation exceptions app. It uses three connectors, QuickBooks, Gmail and Slack.

  • The app holds state. Managed tables store invoice and payment references, match status, exception reason, approvals and run history. This is the memory the model would otherwise carry in a context window.
  • A scheduled workflow triggers the run. A daily Major workflow schedule starts the agent. The trigger is the schedule, not an external webhook.
  • Deterministic code does the routine work. The app matches invoices to payments from QuickBooks by rule, the same way every time. No tokens are spent on clean matches.
  • The agent handles judgment. For ambiguous records it reads related Gmail threads and contract context, then drafts an exception summary.
  • A human approves consequential writes. A Slack approval card goes to the owner before anything changes. Only after approval does the app update the system of record.
  • Governance sits in the app layer. QuickBooks is read-only for the initial fetch, write access is narrowly scoped to approved updates, app access is role-based, and every action is written to the audit log.

The model still does real work here: it judges the exceptions. What leaves the model is the repetition. Costs depend on your plan and on how much of the workflow is ambiguous. For more on this, see the cost model for AI agents. To try this on your own data, start building on Major.

The Major take: price repeatable work once

The constraint with any seat-based builder is multiplication. Roles, tiers, external users, usage and deployment each add a factor, and a headline rate is a weak guide to what recurring internal work actually costs. Every new app also needs a person to build and own it.

Major's answer is narrow and structural. An agent works out a repeatable step, builds a deterministic app for it, and runs that app from then on. Hosting, database and deployment come with the platform, and permissions and audit apply where the agent acts. The model is reserved for judgment. Reason once, run forever.

Start building on Major

Build a stateful reconciliation app that lets an agent resolve routine matches in code and route exceptions for human judgment, with permissions and audit in the app layer.

Related articles

Frequently asked questions

How is Retool pricing calculated?
Retool pricing depends on the plan, the number of builders and internal users, which capabilities sit behind higher tiers, and any usage-based items. Figures change, so confirm current numbers and gates on Retool's official pricing page, then model seats, external access, AI usage and hosting together.
Does Retool charge differently for builders and internal users?
Retool's pricing model distinguishes people who build apps from people who use them, as described on its pricing page and in its pricing v2 explainer. How roles are counted and billed varies by plan and billing period, so verify the current definitions on retool.com/pricing before budgeting.
What should I include when comparing Retool alternatives?
Compare seats by role, external user access, usage-based items such as AI, deployment and hosting ownership, support terms, and which tier holds the security controls you need. Record each as list price, quoted, or unknown, and avoid summing hypothetical rates.
How does Major approach the cost of repeatable agent work?
On Major, an agent builds a deterministic app for a repeatable step once, then runs it as code. Reasoning is front-loaded and the model handles only judgment calls on later runs. Hosting, database and deployment are included. Actual cost still depends on your plan and usage.