Best Internal Tool Builders in 2026: Retool vs Appsmith vs ToolJet vs Budibase — Pricing, Self-Hosting & Real Team Costs

The same ten-builder team with 90 end users pays $0, $1,850, $1,990, or $2,500 a month in 2026 for four platforms that produce the same admin panel — and the $0 option carries no license fees at all, while the $2,500 one is the entry ticket to SSO. That spread, not feature grids, is what actually decides the internal tool builders market in 2026. Retool, Appsmith, ToolJet, and Budibase all solve the same problem: an engineer drags a table, a form, and three buttons onto a canvas, wires them to Postgres or an internal API, and ships an ops panel that would have eaten a sprint of React work. I've built dashboards, CRUD screens, and approval workflows on all four — one for a 15-person startup that never paid a cent, one for a support team that outgrew its vendor's free tier in a month, and one self-hosted deployment that runs a warehouse's entire inbound process. The editors are closer to identical than any homepage admits. The invoices are not.

Four billing meters on the same canvas

Every platform here does the core job: a visual editor with 50-80+ components, connections to Postgres, MySQL, MongoDB, REST, and GraphQL, JavaScript queries for glue logic, role-based access, and a preview-to-production path. On editing speed and component polish, the differences are weeks of muscle memory, not quarters of re-platforming.

What separates them is the meter:

• Retool bills builders and internal end users as separate line items, pools AI credits per plan, and meters workflow runs past 5,000 a month.
• Appsmith bills every user a flat $15 — builder or end user, same price — up to 99 users, then jumps to a $2,500 Enterprise floor.
• ToolJet bills builders only. End users are included up to a plan cap (100 on Team), unlimited on Enterprise.
• Budibase bills by actions: every row save, automation step, and agent tool call draws down a monthly bucket, with seats sold as add-ons.

That is the pattern across internal tool builders in 2026: the meter matters more than the editor, because the meter is what compounds as your app succeeds. A five-builder ops team whose dashboards serve 60 account managers can owe anywhere from $0 to $1,000+ a month depending on which meter their traffic trips. Walk through each before you commit.

Retool pricing 2026: split seats, pooled AI credits, an Enterprise gravity well

Retool remains the category's default and its pricing is the most layered. The free plan covers up to 5 users with 250 pooled AI credits monthly, Retool Cloud, and production environments only — no staging, no audit logs, no Git control, and no external users. It is a real plan, not a demo: a two-engineer team can run internal dashboards on it indefinitely.

Team costs $10 per builder per month plus $5 per internal end user, with unlimited users, 1,000 AI credits, and staging environments. Business is the workhorse tier: $50 per builder plus $15 per internal user, and it unlocks audit logs, rich permissions, unlimited environments, and embedded apps for external audiences. External users are metered separately — the first 250 run $8 per user monthly, the next 250 run $6, and beyond 500 the rate drops to $4 with volume pricing available. Workflows execute 500 times a month free, 5,000 on paid tiers, then $75 per additional 5,000 runs.

Two things deserve a hard look. First, the governance ladder: SSO, SCIM, Git source control, and granular roles all sit on Enterprise with custom pricing. A 30-person company that needs SAML will discover that "custom pricing" in this category means a five-figure annual conversation. Second, self-hosting: Retool lists self-hosted deployment on Enterprise for new customers. Cloud-only is a non-starter for teams with data-residency clauses, and Retool knows its retention rate makes that acceptable. The AI credit pool — 250 to 10,000 per month depending on tier, no rollover — is a genuine convenience for generated queries and AI actions, but it renews monthly and burns fast once your team leans on it.

My honest read after auditing two Retool invoices: the builder-versus-end-user split is fair in principle — builders consume more value than viewers — but the Enterprise gravity well is real. The moment you need SSO, self-hosting, or Git-backed workflows, the published prices stop applying.

Appsmith: open source, Apache-2.0, and the 99-user cliff

Appsmith is the open source option most teams compare against Retool, and the license matters: Apache-2.0, fully self-hostable on Docker or Kubernetes for free, with the same JavaScript-first builder as the cloud. The Community tier is the genuinely free path — unlimited apps, unlimited workspaces, no seat countdown if you self-host.

The cloud lineup is short. Free covers 5 users. Business is $15 per user per month, counting every user identically — a builder and someone who clicks one dashboard twice a week cost the same — capped at 99 users. Business adds audit logs and granular access control. Enterprise starts at $2,500 per month for 100 users, and that is where SAML and OIDC SSO live; Google and GitHub OAuth are available below it. That cliff deserves attention: a 100-person company cannot buy Appsmith Business, full stop. The jump from 99 users to Enterprise is a 17x price step on paper.

Two 2026 changes shape the evaluation. Appsmith's standalone AI feature set ended September 30, 2026 — what remains is Ask AI for JavaScript and SQL snippets on self-hosted instances with your own model key, plus an opt-in MCP beta for coding agents on Business Cloud and self-hosted Enterprise. That is a narrower AI story than Retool's credit-metered assistant or ToolJet's whole-app generation. What you get instead is predictability: no usage meters, no action buckets, no AI credit pools. A team of 40 that builds 20 apps and uses them daily pays $600 a month, and that number does not move when traffic spikes.

Where Appsmith loses deals is polish and breadth: fewer pre-built integrations than Retool's catalog, and an admin experience that assumes someone on your team enjoys maintaining Docker images.

ToolJet pricing: builders only, end users included

ToolJet is the newest serious entrant — AGPL-3.0 licensed, 41,000+ GitHub stars, and a pricing model aimed directly at Retool's weakest flank. It bills builders only. End users ride free up to the plan's cap: 50 on Free and Basic, 100 on Team, unlimited on Enterprise.

The published lineup: Free gives 2 builders, 50 end users, 10 apps, and 100 AI credits monthly. Basic runs $23 per builder monthly on annual billing ($29 month-to-month) with 2,000 AI credits per builder. Pro is $79 ($99 monthly) and adds version control and custom styling. Team is $199 per builder ($249 monthly) and includes SSO via SAML, OIDC, and LDAP, audit logs, Git sync, white labeling, custom domains, and multi-environment support on self-hosted deployments. Enterprise is custom with unlimited end users, SCIM, and air-gapped deployment as an add-on.

Run the same ten-builder, ninety-user math that made Retool's Business tier cost $1,850: ToolJet Team is $1,990 — comparable, but with SSO, audit logs, and Git sync included rather than quoted, and the hundred-user cap is the ceiling, not the start of a different contract. Below ten builders the gap widens fast, because end users never appear on the invoice at all. A five-builder support org with 80 daily users pays $995 on ToolJet Team versus $1,450 on Retool Business with no SSO included.

The differentiator nobody advertises honestly is the AI workflow. ToolJet generates whole apps from a prompt, its MCP server is MIT-licensed so Claude Code, Codex, or Cursor can compose native ToolJet apps against your data without consuming ToolJet credits, and the visual editor remains the fallback for edits. The catch: Workflows and the agent builder are self-hosted only — cloud customers get the app builder but not the automation layer. If your team lives in the cloud plan and needs scheduled automations, budget for a self-hosted instance or look at dedicated automation tooling; our workflow automation comparison covers that separate layer.

Budibase pricing: actions, not seats — with hard limits

Budibase rebuilt its pricing around actions, and it is the most different model of the four. An action is any unit of work: a row save or delete, an automation step execution, an AI agent tool call. Custom SQL queries and API requests are explicitly excluded, which keeps the meter tied to end-user activity rather than engineering cleverness.

The cloud tiers: Pro at $19 a month buys 2,000 actions, one creator, one workspace, and 1-day log retention. Premium at $49 raises the bucket to 20,000 actions, adds 10 workspaces, custom branding, backups, and SSO. Business at $299 covers 250,000 actions, three creators, 30-day logs, enforced SSO, and user groups. Enterprise is custom, with SCIM, audit logs, 365-day retention, and optional air-gapped deployment. End users are a $5-per-user add-on on every tier; extra creators run $50 each. Action limits are hard caps, not overage billing — hit the ceiling and apps stop executing automations until you upgrade.

The free path is the old Budibase virtue intact: a fully open-source self-hosted plan, free indefinitely, no license fees, no action caps beyond your own infrastructure. For a team comfortable running one container, Budibase self-hosted is the closest thing to a $0 total cost of ownership in the category — the trade is owning uptime, upgrades, and backups yourself.

The honest caution: 2,000 actions on Pro is a small bucket. A single busy approval form used by 30 people — submissions, status updates, notification steps — can clear it in days. Budibase's math works best for teams whose apps are light, whose automations are few, or who self-host and sidestep the meter entirely. Heavy multi-app clouds will find Business at $299 is the realistic floor, which lands between Appsmith's Business ceiling and ToolJet's Team tier.

Which internal tool builders fit which team

After building on all four, here is how the low-code internal tools decision shakes out:

• Startup under 5 users: Retool Free or Appsmith Community self-hosted. Both are real platforms at $0, and neither traps you — apps export, data stays in your database.
• Open source mandate or data residency: Appsmith (Apache-2.0) if you want zero licensing friction, Budibase self-hosted if your apps are form-and-automation shaped, ToolJet if you want prompt-built apps with an MIT-licensed MCP server. Retool is out of the conversation for new self-hosted deployments below Enterprise.
• Large end-user audience, few builders: ToolJet, uncontested. End-user-inclusive billing is the single biggest cost advantage in the category once your apps get adopted.
• Flat, predictable cost is a procurement requirement: Appsmith Business — $15 a head, no usage meters, no credit pools — as long as you stay under 100 users.
• Existing Retool estate or mobile apps: stay on Retool. Migration costs more than the per-seat delta, and the workflow runner plus agent features are the most mature here.
• Form-heavy, automation-light internal apps: Budibase. Blocks, automations, and the built-in database make simple CRUD fastest to ship — just model your action volume before choosing a tier.

The comparison table

Every figure below comes from the vendors' official pricing pages as of October 2026. This category reprices quietly — Budibase's switch to actions is the third model change among these four in two years — so verify before contracting.

Frequently Asked Questions

Retool vs Appsmith: which should a startup actually pick?

Under 5 users, it hardly matters — both free tiers run real production apps. Past that, the decision is billing shape: Appsmith charges $15 for every user identically and stops at 99, while Retool charges builders more than viewers and scales smoothly past 100 users on published prices. A 10-person engineering team serving 50 ops users pays $300 on Retool Team versus $900 on Appsmith Business; the same team serving 150 users pays $850 on Retool versus an Enterprise contract on Appsmith. Pick by your user ratio, not by ideology — and note Retool's SSO still requires Enterprise either way.

What is the best open source Retool alternative in 2026?

Appsmith for JavaScript-first teams that want the largest community and Apache-2.0 licensing with zero strings; ToolJet for teams that want whole-app AI generation, an MIT-licensed MCP server for coding agents, and a published per-builder price that includes end users; Budibase for form-and-automation apps that can live entirely on its built-in database. All three self-host free on Docker or Kubernetes. The one thing none of them gives you is Retool's integration catalog — budget engineering time to wire niche internal APIs yourself.

Is Retool worth it, or are Retool alternatives good enough?

Retool is worth it in three cases: you already run Retool apps, you need its mobile offering or workflow maturity, or you are an enterprise where the custom quote is procurement overhead rather than a risk. Everywhere else, the alternatives are good enough on the canvas and materially better on the invoice. My contrarian take after auditing real bills: Retool's Enterprise gravity well is the product strategy, not an accident. SSO, self-hosting, and Git control sit a tier above where teams expect them, and that is where the published pricing ends.

ToolJet pricing for a small team: what will we really pay?

Two builders and modest usage: $0. Five builders building actively: $115 a month on Basic annual — and every end user up to 50 rides free. The number that surprises teams is the jump to Team at $199 per builder, which you only need when SSO, audit logs, or Git sync enter the picture — at ten builders that is $1,990 monthly, comparable to Retool Business but with governance included.

Can I build internal tools for free forever?

Yes, three ways. Self-host Appsmith, ToolJet, or Budibase — all open source — on infrastructure you already pay for; the software license costs nothing and the only cost is an engineer's maintenance time. Or stay under the cloud free tiers: Retool's 5-user plan, Appsmith's 5-user cloud, ToolJet's 2-builder plan. The realistic ceiling of free infrastructure in 2026 is a team of five running ten apps with light traffic — enough for most early-stage ops needs. When AI-generated apps and agent-built workflows tempt you to build more, that is the point to read our AI agent frameworks comparison and decide whether you outgrew drag-and-drop.

The Bottom Line

All four internal tool builders will produce the same admin panel in an afternoon — the editor is a solved problem, and the billing model is the product. Retool wins on maturity: the deepest integration catalog, the most polished workflow runner, and mobile apps no open source rival matches. Appsmith wins on honesty: Apache-2.0, a flat $15 with no hidden meters, and a self-host path that never expires. ToolJet wins on adoption economics — end users included, SSO and audit logs at a published price, and a coding-agent story that turns Claude Code into an app builder. Budibase wins on the $0 self-host floor and the cleanest action-metered cloud for form-heavy apps.

My own setup after this comparison: Appsmith self-hosted for the warehouse deployment with data-residency clauses, Retool Free for the client project nobody wants to migrate, and ToolJet Team for the support org where 90 daily users would have added $1,350 a month to a Retool invoice. Total platform spend: one VPS and $995 a month. In 2026 the canvas is free — you are only ever shopping for a meter, so pick the one your traffic doesn't trip.