Opportunities › Desktop automation via OS accessibility APIs › Proposed app
Casement
Turn any desktop app into an API your scripts and AI agents can call
- Score
- 39
- confidence 47
- Evidence
- 4
- items on the page
- Rivals
- 0
- named in sources
- First version
- 45 d
- estimate
- Type
- Desktop app
- PWA: runs in the browser, installs on Mac and Windows
- Market since
- 2006
- Mature market · AI estimate
ScreensConcept
Desktop app: Engineers and developers work at their computer to configure automation, inspect accessibility trees, debug workflows, and manage API calls with precision. Sample screens on a computer, in the proposed brand. Their numbers are sample data.
Element tree
Bridge online42 elements
About this appAI proposal
Casement reads the accessibility tree that Windows, macOS and Linux already expose and turns each window into a named set of controls. A small local bridge runs on the machine; you call buttons, fields and checkboxes over REST or MCP instead of writing a script per application.
It is for engineers, integration specialists and AI developers who need to drive existing desktop software without reverse-engineering it. The PWA console is where you browse the live tree, try an action, save it as a named endpoint, and review every call your agents made.
What is different: calls are saved as named actions with a tested selector and a replay log, so a changed layout shows up as a failing check you can read, not a silent break in a recorded macro.
- Tagline
- Open a window on any desktop app, and call it like an API.
- Voice
- PlainPreciseCalm
- Why this name
- A casement is a window that opens outward on a hinge. The product opens windows of existing apps so code can reach in. Short, plain and easy to say.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Live element tree
Browse any open window as a tree of buttons, fields and checkboxes read from the OS accessibility layer. Select a node to see its role, name and state.
No member yet
Try it, then save it
Send a click, type or read from the console and watch it happen. Save the working call as a named action with its own REST URL.
No member yet
MCP server built in
Expose saved actions as MCP tools so coding assistants and agents can use desktop apps with typed inputs and clear descriptions.
No member yet
Selector health checks
Run saved actions on a schedule against a test machine. Learn which selector stopped matching after an app update.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#1F5E55Pine · primary#F4F1EALinen · surface#1F2A28Ink · text#D9822BAmber · accent#3B6EA5Slate Blue · data
Typography
Aa
Casement
ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
Open a window on any desktop app, and call it like an API.
PricingAI proposal
The material gives no competitor prices. These tiers are a starting point to test, based on the monetization idea of subscriptions for volume, support and a hosted service.
Free
Free
- One machine
- Element tree and try-it console
- Local REST and MCP
- 7-day call log
Pro
$29 a month
- Five machines
- Saved actions and health checks
- Approvals and scoped keys
- 90-day call log
- Email support
Team
$99 a month
- Twenty machines
- Shared action library
- Hosted relay for remote machines
What it could be worthEstimate
What a software business making this much sells for today, from real deals. Pick a monthly revenue.
If it makes
Worth about
$194,000
Sale range $132,000 to $263,000
$120,000 a year in revenue
From 87 software businesses listed and sold on Empire Flippers in the last year, revenue vetted by the broker: they sell for about 19× their monthly revenue, most between 13× and 26×.
Measured on profit, the middle deal is 32× monthly net profit: about 2.7 years of profit.
An estimate, not an offer or financial advice. A real price depends on profit, growth, churn, how much of the work is the owner's, and the buyer.
How it would growAI proposal
How the product would reach new customers through its own use, and how the first ones would be found. A plan to test, not a result.
- Built-in loop
- No loop in the product. Casement runs on a developer's own machines and behind their agents. Nothing it makes is shown to other people, so growth comes from content, open examples and community posts rather than from use.
- Do they become customers?
- Reach comes from sales and content. Test whether developers who find an example or MCP listing install the bridge and then pay for saved actions, health checks or approvals.
- First customers
- Integration engineers at small firms that must automate one legacy Windows desktop app with no API, such as accounting or back-office software.
- Where to reach them
- Hacker News and developer forums: Share a worked example of driving a legacy app through MCP, and invite people to try the free tier.
- MCP and AI-agent communities: Publish a ready MCP server and sample actions in agent tool directories and Discord groups.
- Direct outreach to integrators: Offer to build one automation with five integration consultancies and learn what they would pay for.
- First milestone
- Five paying Pro users, each running health-checked actions against a real legacy app for 30 days.
Why people already pay for this
- VerifiedHacker News post received 18 points and 9 comments
- Public claimOculOS converts desktop UIs to REST-controlled interfaces by reading OS accessibility trees, enabling programmatic control of buttons, text input, and checkboxes.
- Public claimThe tool integrates with AI coding assistants through MCP protocol, enabling them to control desktop apps natively.
- Public claimThe solution is lightweight and platform-agnostic, supporting Windows, Linux, and macOS.
Tech stack
Every proposal builds on the same PWA base, installable from the browser on any phone or computer. Dashed boxes are what this app adds.
- AppNext.js and TypeScript
- PWAManifest and service worker
- UITailwind CSS
- Data and authSupabase Postgres
- PaymentsStripe
- EmailResend
- HostingVercel
- This app · Local bridgeSmall cross-platform native agentA PWA cannot read other apps' accessibility trees. A bridge on each machine must call the OS accessibility APIs.
- This app · Agent protocolMCP server and REST gatewayLets coding assistants and scripts call saved actions in the form they already use.
RisksAI inference
- TechnicalAccessibility trees differ by app and toolkit. Some apps, such as games and custom-drawn UIs, expose few or badly named elements, so coverage will be uneven.
- PlatformmacOS and Windows require users to grant accessibility permission and may tighten it. Signing and installing a native bridge adds friction for a PWA product.
- CommercialThe source is one Hacker News post with 18 points and 9 comments, and open-source tools do similar work. Willingness to pay for a console is unproven.
- RegulatoryAn agent that can click anything on a machine can touch sensitive data. Buyers will expect scoping, approvals and audit logs, and some may require security review.
- AIAgents may misread controls or act on the wrong element. Without approvals and replay, errors in billing or records could be costly.
Everything marked AI proposal or AI inference was written by a model from the evidence on the opportunity page. The proposed name must be checked against trademark registers before it is used.