From a Failed Fintech to a $3.2B Enterprise Platform
Before Retool existed, David Hsu was trying to build something else entirely. A fintech startup called Cashew - gone. What remained was a very specific frustration: every startup he'd touched spent enormous engineering effort building the same boring internal tools. Admin panels. Dashboards. CRUD apps. Things that were never the product, but always required precious product time.
Hsu took that frustration and, alongside Anthony Guo - then fresh out of Palantir, a company that practically invented the high-end internal data tool - turned it into a company. Retool launched quietly in 2017 out of Y Combinator's W17 batch. By the time they were ready to announce to the world, they'd already hit $2 million in annual recurring revenue. Before the press release. Before the Product Hunt launch. Before most people had even heard the name.
That detail says a lot about how Retool operates. They didn't pitch a vision and wait for traction - they found developers who had a real problem, sold them a real solution, and let the product do the talking. The "try" in their GitHub handle - tryretool - is not accidental. It's the whole strategy.
They reached $2M ARR before officially launching publicly - by quietly selling to developers who actually needed internal tools.
Hsu's background is unusual for a Silicon Valley CEO: he studied philosophy and computer science at Oxford. That combination shows up in how Retool thinks about product. It's not just about what the tool does - it's about how developers think, and how to get out of their way. Guo's time at Palantir, meanwhile, gave the company a clear picture of what enterprise-grade internal tooling actually requires at scale.
The Platform That Builds the Software Nobody Sees
Most software companies want to build the thing customers use. Retool builds the thing that runs everything else - the internal dashboards, ops tools, customer portals, and admin panels that keep companies working. It is not glamorous. It is, however, genuinely essential.
The core product is a drag-and-drop builder. Connect it to a Postgres database, a REST API, or a third-party service. Drop in a table, a form, a chart. Write a bit of JavaScript if you need custom logic. Deploy in minutes. For a developer who's stared at a weekend of boilerplate ahead of them, it's a significant relief.
What made Retool different from earlier no-code tools was its audience. This was not built for non-technical business users who want to "build apps without code." It was built for developers who are tired of building the same app for the fifth time. The interface is flexible enough for engineers but not so raw that it requires deep expertise. That gap - between spreadsheets and full custom development - is where Retool lives.
In 2025, the company pushed aggressively into AI. Retool Agents arrived in May 2025, promising to automate complex workflows through autonomous AI agents. Within months, the platform was reporting 100 million hours of work automated. Retool AppGen - which generates full apps from a natural language description - went generally available in October 2025. The pitch is shifting: it's no longer just "build internal tools faster." It's "let AI build them for you, then deploy agents to run them."
Six Tools. One Platform.
App Builder
The original. Drag-and-drop interface for building web apps on top of any database or API. Tables, forms, charts, buttons - all wired up to real data.
AppGen
Describe your app in plain English. Retool's AI generates the full thing. GA'd October 2025. The bet: non-engineers shouldn't need to learn the builder to use Retool.
Agents
Build and deploy autonomous AI agents that handle multi-step workflows on their own. 100M hours automated since May 2025 launch. The most talked-about new direction.
Workflows
Visual automation for multi-step processes. Chain APIs, databases, and internal systems together without writing a full backend service.
Mobile
The same builder, but for native iOS and Android apps. Operations teams get mobile tools without a separate mobile development budget.
Database
A built-in PostgreSQL database so teams don't need a separate data store to start. Keeps the "build and deploy" cycle genuinely fast.
From Scrappy Startups to Fortune 500 Back Offices
Retool's customer list reads like a roll call of companies that move fast and have no patience for slow internal tooling. DoorDash uses it. Amazon does. So does Mercedes-Benz, NBC, Brex, and Plaid. The common thread is not size or industry - it's that they all have developers who need to build operational software quickly and don't have time to build it from scratch.
Over 500,000 apps have been built on the platform. That's a remarkable number for a category - internal tools - that by definition never gets announced at a product launch. These are the dashboards that operations teams live in all day. The admin panels that customer support opens when something goes wrong. The tools that make companies work, even if nobody outside the company will ever see them.
Retool's pricing reflects its audience. There's a free tier for developers kicking the tires. Standard plans run $12 per user per month. Business plans at $65. Enterprise is negotiated. Viewers - people who consume the app without editing it - are free, which is a smart move: it lowers the barrier to deploying Retool across an org once the builders are already sold.
The Macro Trend Nobody Was Writing About
In February 2026, Retool published its annual Build vs. Buy Report with a finding that landed with some weight in enterprise circles: 35% of enterprises had already replaced SaaS software with custom internal tools. Not building alongside their existing SaaS - replacing it.
It's a trend Retool has been quietly enabling for years, but the numbers make the scale concrete. Off-the-shelf software is expensive, inflexible, and increasingly being undercut by AI-assisted development tools that make custom software more accessible. Retool's bet - that every company eventually wants software built specifically for how they operate, not how a software vendor imagined they might - is looking increasingly correct.
The AWS strategic collaboration (December 2025) and the Databricks Partner of the Year award (June 2025) signal something specific: Retool is not trying to be a standalone platform in isolation. It's threading itself into the infrastructure enterprises already depend on. That is a different - and arguably more durable - growth strategy than trying to be yet another self-contained SaaS product.