1. Describe it to your AI
Ask in plain language through your tenant's MCP server ({tenant}.plaid.cloud/mcp), from a Claude connector or any MCP-capable AI. No boilerplate and no infrastructure to wire up.
Build AI Apps on Your Data
Describe the app you need to your AI. PlaidCloud builds and serves it as a secure HoloViz Panel app, WASM or server-side, running directly on your lakehouse tables, workflows, dimensions, and access controls.
How it works
From description to deployed app
2. PlaidCloud builds & deploys
Your app is built automatically and served as a HoloViz Panel app, WASM in the browser or server-side: versioned, monitored, and ready to share in minutes.
3. It runs on your governed data
The app reads your lakehouse tables, workflows, and dimensions under your existing roles and permissions. Secure by default, at scale.
Stop configuring SaaS. Start describing.
Apps that fit your business exactly
Off-the-shelf SaaS apps force your process into their template. On PlaidCloud, your people describe the app they actually need and get it in less time than a SaaS configuration project, running on your data instead of another vendor's copy of it.
- No new vendor, contract, or data copy per app
- Apps read live, governed data instead of exports
- Revise an app by asking the AI, not by filing tickets
Governance is built in, not bolted on
Your roles and permissions apply to every app
Every app runs inside your tenant under the same roles, permissions, and lineage as the rest of PlaidCloud. Data never leaves your governed platform, and access can be as broad or as narrow as the audience requires.
- Apps inherit workspace, project, and user-level access controls
- Versioned and monitored as managed server apps
- Traceable back to the tables and workflows they read
Built on open-source HoloViz Panel
Two ways to serve: WASM or server-side
Apps are standard HoloViz Panel applications, built and hosted by PlaidCloud. Serve them as WASM so the app runs entirely in the user's browser, or server-side when the app should run next to the data. Either way, PlaidCloud's automated build process takes care of everything: no local development environment, no Python setup, no build tooling.
- Open-source HoloViz Panel, not a proprietary app framework
- WASM serving: the app runs in the browser, nothing to size or scale
- Server-side serving: managed apps running next to your data
- Automated builds: create and deploy without a local dev environment
Your apps live in Git
Provided repos, or bring your own GitHub / GitLab
Every Panel app is backed by a Git repository. Use the repos PlaidCloud provides, or connect your own GitHub or GitLab. AI edits land as commits, so every app has reviewable history and rollback like any other codebase.
- PlaidCloud-provided Git repos out of the box
- Or bring your own GitHub or GitLab
- AI-built and AI-edited apps are versioned like any codebase
What teams build
Example apps, described in a sentence
Margin review app
A weekly margin-movement review for sales leaders, reading live allocated P&L from the lakehouse.
Close checklist tracker
A month-end close tracker that watches workflow run status and flags late or failed steps.
Plant cost explorer
A self-service explorer for plant controllers to drill from site cost to the transactions behind it.
Driver input collector
A governed form that collects allocation driver updates from the business and lands them in model tables.
Frequently Asked Questions
How do I build an app with AI on PlaidCloud?
Point your AI at your tenant's MCP server, describe the app you need in plain language, and PlaidCloud builds and serves it as a HoloViz Panel app. There is no local development environment, hosting, or build tooling to set up.
What is a HoloViz Panel app?
Panel is a leading open-source Python framework for building interactive data apps: dashboards with inputs, forms, drill-downs, and custom logic. PlaidCloud apps are standard Panel applications, so they are portable Python, not a proprietary format.
Can PlaidCloud host Panel apps?
Yes. PlaidCloud builds and hosts HoloViz Panel apps end to end, served either as WASM running in the browser or server-side next to your data, with versioning, monitoring, and your access controls applied.
What is the difference between WASM and server-side serving?
WASM apps run entirely in the browser, so there is no server to size and they scale effortlessly for broad audiences. Server-side apps run inside your tenant next to the data, which suits heavier computation and tighter data access. PlaidCloud builds and serves both.
Do I need a local development environment?
No. PlaidCloud runs an automated build process, so you can create, edit, and deploy Panel apps entirely from the platform, including with AI. Nothing to install locally.
Where does the app code live? Can I use my own Git?
Every app is backed by a Git repository. PlaidCloud provides repos out of the box, or you can bring your own GitHub or GitLab. AI edits arrive as commits, so history is reviewable and revertible.
What is the per-tenant MCP server?
Every PlaidCloud tenant includes its own MCP (Model Context Protocol) server at {tenant}.plaid.cloud/mcp. It exposes your tables, workflows, dimensions, and app deployment to any MCP-capable AI, such as a Claude connector, under your access controls.
Are the apps secure?
Yes. Apps run inside your tenant and read data under your existing roles and permissions. Nothing is exposed publicly unless you choose to share it, and access is controlled the same way as the rest of PlaidCloud.
Build your first app this week.
Bring one real need to a 30-minute session and leave with a deployed app on your own data.