Ryz Labs/Replace/Replace Retool
Replace

Replace Retool with Custom Internal Tools: Runbook

A renewal-season runbook for Retool: what the builder and internal-user seats on your bill add up to, where owned code crosses over, and how to move apps off safely.

Replacing Retool with custom internal tools makes sense when internal-user seats grow with headcount, the apps run core operations, and you have engineers who will own the code afterward; it rarely makes sense for a few dozen users on standard admin screens. The number that decides it: builder seats × builder price plus internal-user seats × internal-user price, over 36 months, against engineers × months × $7,000 plus a yearly run cost. In our worked example on Retool Business list prices, a 300-user org crosses over in month 24, and 25- and 100-user orgs never do.

What you pay for today

Retool bills two kinds of internal seats at different prices, so most bills are a mix. On the Retool pricing page, Business is listed at $50 per builder and $15 per internal user per month billed annually (list price checked November 21, 2026; check your contract). A Business org with 10 builders and 90 internal users lists at 10 × $50 + 90 × $15 = $1,850 per month.

PlanList priceUnitNotes
Free$0Up to 5 users500 workflow runs per month
Team$10 builder, $5 internal user billed annually; $12 and $7 billed monthlyPer user per month5,000 workflow runs per month
Business$50 builder, $15 internal user billed annually; $65 and $18 billed monthlyPer user per month5,000 workflow runs per month
Business external usersFirst 50 free; 51–250 at $8, 251–500 at $6, over 500 at $4 billed annually ($10, $7, $5 monthly)Per external user per monthTiered by external-user count
EnterprisePricing is by quote (contact sales)QuoteSelf-hosted deployment is Enterprise only

List prices checked November 21, 2026. Check your contract; negotiated prices differ.

Usage sits on top of seats. Team and Business include 5,000 workflow runs per month, and the pricing page shows no overage price. AI credits range from 250 to 10,000 per month depending on plan; credit packs are sold, but the page doesn't price them. Agents get 20 free hours per month, then bill at a per-model hourly rate that the page points to the docs for.

What moves the bill at renewal, stated plainly:

When to keep Retool, and when to own it

Retool is a strong product for what it was built to do: putting screens on top of databases and APIs fast. Many teams should keep it. The question is whether a specific set of apps has become software your business runs on.

SituationRecommendation
Under about 25 users, mostly CRUD and admin screens over databases you already ownKeep Retool. At these seat counts the run cost of owned code alone exceeds the bill.
No engineer can own the apps after launchKeep Retool. Owned code without an owner decays faster than a vendor contract.
Builders ship many short-lived apps a quarter (one-off reports, migration consoles)Keep Retool. Fast throwaway screens are its strength.
You rely heavily on Retool Workflows, AI credits and agents across many appsKeep, or price owned replacements one workflow at a time.
Internal-user seats grow with every hire in operationsConsider owning the apps those users live in.
An app is core operations (an underwriting queue, a fulfillment console) and its UX or logic differentiates youConsider owning it, with tests and code review like the rest of your product.
Data lives in Retool Database and must sit in your own databaseMove the data to your PostgreSQL first; you can keep the apps on Retool while you decide.
Security requires self-hostingCompare a Retool Enterprise self-hosted quote against owning; both are real options.
Middle pathKeep Retool for builders and long-tail admin apps; move the two or three high-traffic apps that drive internal-user seats to owned code, and shrink the internal-user count at renewal.

3-year break-even math

Renting side: add builders × $50 and internal users × $15 for one month, then multiply by 36. Owning side: engineer-months at $7,000, plus the Retool subscription you can't drop until month 5 (four build months and an overlap month), plus hosting and upkeep starting in the overlap. We mark break-even at the first month where owning's running total has caught up with renting's and stays caught up for good.

What the rows assume (change any of it):

Seats (builders + internal users)List cost per year3-year list costBuild teamAssumed run cost per year3-year owned costBreak-even month
25 (5 + 20): $550 per month$6,600$19,8002 × 4 months × $7,000 = $56,000$24,000$122,750Not reached: $2,000 upkeep beats a $550 bill
100 (10 + 90): $1,850 per month$22,200$66,6002 × 4 months × $7,000 = $56,000$24,000$129,250Not reached: $2,000 upkeep beats a $1,850 bill
300 (20 + 280): $5,200 per month$62,400$187,2002 × 4 months × $7,000 = $56,000$24,000$146,000Month 24

Seat math on Retool Business annual prices seen November 21, 2026 (list price, check your contract); engineers priced at $7,000 a month; upkeep assumed.

How the 300-user row works: 3-year owned cost is 5 months of Retool (5 × $5,200 = $26,000) + $56,000 build + 32 months of run cost (32 × $2,000 = $64,000) = $146,000. At month 23, owned is $120,000 against $119,600 of list; at month 24, owned is $122,000 against $124,800, and the gap widens by $3,200 a month after that. For the 25- and 100-user rows the monthly Retool bill is below the assumed $2,000 run cost, so owned code never catches up, however cheap the build. At those sizes, keep Retool or take the middle path.

Your real numbers will differ: negotiated prices, a different seat mix, the Team plan, or a larger app estate. Our renewal-or-build calculator takes your actual quote. Treat the rows as a picture of where two cumulative curves intersect, not a forecast of savings.

What a Retool replacement involves, step by step

1. Inventory apps, users and resources

List every app and workflow, who owns it, which groups use it and which resources (databases, APIs) it queries. The Retool API takes a Bearer token that an org admin creates under Settings, Retool API, with scopes for users, groups, folders and permissions. Its rate limit is points-based: up to 300 points in a 60-second window, where apps, folders and users cost 2 points, groups, resources and source control 5, and permissions 10. Going over blocks calls for 60 seconds, so pace permission pulls on a large org.

Exit check: a sheet with one row per app: owner, user groups, resources, internal-user count and a keep, move or retire decision.

2. Export app definitions into a repo

Retool's import and export docs describe exporting an app as a single JSON file or a Toolscript ZIP; the export contains its queries, components and modules. Toolscript lays an app out as main.rsx, functions.rsx and metadata.json, with SQL and JavaScript extracted into a lib folder. Those lib files are your best starting point: each SQL query becomes a parameterized query in your API, and each JavaScript transformer becomes a tested function. YAML serialization was removed in the Q2 2025 stable release, and Toolscript is available on cloud and on self-hosted 3.6.0 and later. If your org uses Source Control, apps, workflows, queries, resources and themes already sync to GitHub, GitLab, AWS CodeCommit, Bitbucket or Azure Repos through pull requests. Workflows export separately as JSON with their blocks and queries.

Exit check: every app and workflow you plan to move is in a repo, and every query in lib is tagged with the resource it runs against.

3. Recreate resources and settle the data model

Resources are not in app or workflow exports, so connection details live only in Retool until you rebuild them in your backend with your own secrets manager. Many Retool apps already query your databases, so this step is usually lighter than for tools that hold your records. The exception is Retool Database: cloud orgs get the equivalent of 5 GB of PostgreSQL data with unlimited tables and rows, backed up daily and kept for 7 days, and Retool's Retool Database docs say it isn't built to be highly available or scalable. Retool's documented path for moving that data, in its cloud-to-self-hosted guide, is exporting each table as CSV and importing it. Load those tables into your PostgreSQL with real constraints and migrations in version control.

Exit check: row counts and checksums match per table, and no app depends on a resource that exists only in Retool.

4. Rebuild the seat-heavy apps first

Start with the app most internal users open daily, because it moves the bill and gets the most testing. A typical stack is a React and TypeScript front end, an API in Node.js or Python, and your PostgreSQL. Map Retool groups to roles in your app. Record real inputs and outputs from the Retool version before you start, and turn them into tests.

Exit check: every recorded case passes on the owned app, it is gated by a feature flag, and someone is on call for it.

5. Move workflows, integrations and auth

Retool workflows become scheduled jobs or queue consumers in your codebase. Retool's own cloud-to-self-hosted guide warns that workflow webhook URLs change during that migration; moving to owned code changes them too, so find every system that calls a Retool webhook. Put the apps behind your identity provider with SSO and write an audit log of who changed what.

Exit check: every webhook caller is repointed, SSO is enforced, and audit events land where your security team reads them.

6. Overlap month

Run both versions for at least one billing month. Move users group by group, compare outputs from the same inputs daily, and keep Retool as the fallback until the diff is empty for a full cycle of month-end work.

Exit check: zero unexplained output differences and no user still opening the old app for a task the new one covers.

7. Cutover and decommission

Remove migrated users from Retool before the renewal date so the next term is priced on the seats you still need. Archive the final JSON or Toolscript exports in the repo. Before canceling anything, confirm in writing how long Retool keeps your org's data after cancellation, and export Retool Database tables one last time.

Exit check: seat counts reflect the move, final exports are archived, and the retention answer is filed with the contract.

Replacement checklist

What an AI pod on your team does in the first two weeks

Week one

Week two

Assistants write the first draft of the dull pieces for our AI pod teams: typed endpoints from extracted SQL, tests from recorded app sessions, export and pacing scripts. People still review each pull request, and tests gate every merge.

Typical team shape

The shape we see most for a Retool exit: a tech lead who owns the app sheet and decides migration order; one to three senior full-stack engineers who know React and TypeScript on the front and Node.js or Python on the back; a data engineer only when Retool Database holds a lot; and QA a few days a week while both versions run. Many teams begin with a pod of 3 to 5 senior engineers (a tech lead plus engineers), and a full pod is about seven. A smaller replacement can start with a tech lead plus one or two engineers.

Ryz engineers on your team start from $7,000 per month (our published range is $7,000–$15,000 per engineer per month, mostly mid-level to senior), quoted per team. Once you say go, engineers can typically start in under two weeks. Ryz engineers work on your team, reporting to your leads. Send us the app sheet and we'll scope your team.

Related

FAQ

Is it cheaper to build than to renew Retool?

Only past a certain headcount. Price your builders and internal users separately, total 36 months, and compare with engineer-months at $7,000, the Retool months you keep while building and overlapping, and yearly upkeep. In our Business list-price example, 300 seats catch up in month 24; 25 and 100 seats don't, since upkeep alone outruns their monthly bill.

How do we get our apps and data out of Retool?

Export each app as a JSON file or a Toolscript ZIP, which contains its queries, components and modules, and export workflows as JSON. Resources are not included, so you recreate connections yourself. Data in Retool Database moves by exporting each table as CSV. Data in your own databases stays where it is.

How long does a Retool replacement take?

Count the apps, the JavaScript transformers and workflows inside them, and the resources to reconnect, then add an overlap long enough to include month-end. Two or three busy apps and a full estate of hundreds are not comparable jobs; size the work app by app.

What does a Retool replacement cost?

Multiply engineers by months by rate, and add yearly hosting and maintenance time. Our engineers start from $7,000 a month, and we price the team after seeing your app sheet rather than as a flat figure.

Can we keep Retool for part of the work?

Often that is the right call. Keep Retool for builders and low-traffic admin apps, move the high-traffic apps that drive internal-user seats to owned code, and reduce internal-user seats at renewal. Moving Retool Database tables to your own PostgreSQL is also possible while the apps stay on Retool.

Questions we didn't answer? Email info@ryzlabs.com.

Thanks — your message has been sent. We’ll get back to you soon.

Something went wrong while sending your message. Please try again or email info@ryzlabs.com.

Explore Ryz Labs

Software development servicesAI development companyIndustriesStaff augmentationAI pod teamsNearshore software developmentHire engineers by roleEngineer ratesGuidesBuyer guidesCase studiesHow we vet engineers

Ryz Labs

Senior engineers on your team. AI pod teams that ship.

Tell us what you're building. You get a scoped plan, a price and the names of the people who would do the work.

  • Only the top 1% of tens of thousands interviewed make it
  • On US business hours, including New York hours
  • Trusted by Fortune 500 engineering teams

Tell us who you need

A Ryz partner replies with a scoped team plan.

Start a conversation →

Prefer to talk? Book a 15-min call →