clundra

Designed for builders. Loved by agents.

Build software.
Share it like a document.

Bring a repository, a container image, or both. Clundra connects them into one secure application — one URL, access you control, deployments that happen on their own, and a way back from every change.

See how it works
  • One app. One URL.
  • Private services.
  • Built-in access.
  • Rewind any deploy.
FrontendGitHub repo
BackendContainer image
DatabasePostgres
customer-portal
customer-portal.clundra.app
Running
v25Deploying
v24Working
v25Health check failed
v24Rewound
Share access
Invited people
Anyone with the link
NnicuOwner
FfinanceCan edit

One app, not a cloud project

Run software.
Not infrastructure.

Add the pieces your app needs. Clundra works out how they build, connect, deploy and recover. You end up with one application — not a drawer full of cloud resources you have to remember.

New app
Add what it needs. Clundra connects the rest.
App name
customer-portal
FrontendGitHub · acme/customer-portal-web
BackendImage · ghcr.io/acme/api:latest
PostgreSQLManaged · 1 GB
Advanced settings
Deploy app →

Placeholder — swap for a real console capture

customer-portalcustomer-portal.clundra.app
LiveShare
FrontendBackendPostgreSQL
Recent deployments
v24  Healthy2 min ago
v23  Rewound8 min ago
Share “customer-portal”
Who can open it?
Only me
Invited people
Entire workspace
Anyone with the link
NicuOwner
FinanceViewer
AnaEditor
Copy app link

Shareable by default

Share an app
like a document.

Keep it to yourself, invite three people, hand it to the whole team, or open it to anyone with the link. Change your mind later without redeploying.

Clundra can authenticate people before traffic ever reaches the app — even when the app itself contains no login code.

Simple first

Start with the defaults.
Open the details when needed.

Clundra works out how services build and connect. The advanced controls stay folded away until your application actually needs them — and they open in place, not on some other page.

Frontendacme/customer-portal-web
GitHub
Backendghcr.io/acme/api:latest
Registry
DatabaseManaged · 1 GB
PostgreSQL
Environment variables
API_URL = /api
NODE_ENV = production
Deployment
Auto deployON
Branchmain
Build
npm run build
npm start
Domains
app.acme.com
customer-portal.clundra.app
Health check
/health
Port & exposure
3000 · Public

Click the row — the settings open in place

Built-in way back

Every change
has a way back.

A bad deployment, a broken setting or an agent’s mistake shouldn’t turn into an incident. See exactly what changed, then return the whole application — services, config, connections — to its last working state.

See recovery in action →
Application history
10:42v21 deployed
All services healthy
11:08v22 deployed by agent
Health check pending
11:10Health check failed
Backend returned 503
11:11Rewound to v21
Application recovered
Human
customer-portal
FrontendBackendPostgreSQL
Deploy new version
Agent
clundra.deploy_app({
  app: "customer-portal",
  verify_health: true,
  rewind_on_failure: true
})
Deployment completed
Health verified
Audit event recorded

Human + agent first

Built for humans.
Operable by agents.

Everything in the dashboard is also in the API. Nothing is dashboard-only, and nothing is API-only.

An agent can deploy, inspect, verify and recover an application without stepping around your permissions — and every action it takes is written to the same history you read.

Production foundations

Small in audience.
Not in importance.

A tool five people use still needs secure access, data that survives a restart, networking that behaves the same tomorrow, and a way back when something breaks.

Clundra ships those foundations with the app, so you don’t need a platform team to have them.

  • Isolated runtime
  • Private service network
  • HTTPS included
  • Managed secrets
  • Persistent storage
  • Deployment history
  • Recoverable states
Internet
HTTPS + access
Frontend
private
Backend
private
PostgreSQL
IsolationSecretsPersistent dataRewind

A cloud for software that doesn’t need a platform team

Internal tools
Dashboards, approval flows, operational utilities, the thing finance asks for every quarter.
Micro-SaaS
Focused products with a frontend, an API and a database. Nothing more.
Agent-built apps
Purpose-built tools created and operated by coding agents, with permissions intact.
Prototypes that stuck
Experiments that now need real data, real access and a way to not fall over.

Small by audience.
Not by value.

Simple enough to start. Real enough to keep.

Starter
Free
Prototypes and personal apps.
  • 1 app
  • Frontend + backend
  • Clundra URL
  • Private access
  • Limited Postgres
  • 7 days of history
Builder
$24/mo
Live internal tools and focused products.
  • Multiple apps
  • Multiple services
  • Custom domains
  • Team access
  • Persistent Postgres
  • Extended history

Placeholder pricing

Availability & roadmap

Available now.
Shipping next.

Apps, sharing and rewind are live today. Here’s what you can use right now — and what’s on the way.

Available now
  • Frontend + backend appsA repo, an image, or both — one application.
  • Managed PostgresConnected and persistent by default.
  • Share like a documentPrivate, invited, workspace, or link.
  • HTTPS + built-in accessAuthentication before traffic reaches the app.
  • Deployment historyEvery change, who made it, what it touched.
  • RewindReturn the whole app to its last working state.
Shipping next
  • Fork from a versionBranch a running app from any point in its history.Q4 2026
  • Team access controlsRoles, scoped permissions, shared workspaces.Q4 2026
  • MachinesThe reversible compute primitive, exposed directly.Q1 2027
  • MCP serverAgents operate Clundra natively, inside your permissions.Q1 2027
  • More regionsRun closer to the people using the app.Q2 2027
  • Deeper observabilityMore insight into state, changes and recovery.Q2 2027
NowNext

FAQ

Questions before you deploy?
Here’s what teams ask most.

Yes, for the software this is built for: internal tools, focused products and small services. You get isolation, private networking, HTTPS, managed secrets and persistent data on every app — not just the paid tiers.

The whole application returns to a previous version — every service, its configuration and the connections between them, together. Persistent data tied to that point comes back with it. Nothing to reconfigure, and the rewind itself is recorded in the history.

No. Clundra can authenticate people at the edge, before traffic reaches your app, so a prototype with no auth at all can still be shared safely with three named colleagues.

Everything in the dashboard is in the API, so yes — with the same permissions you have and the same audit trail. A native MCP server is on the roadmap; today it’s the REST API.

If it runs from a repository or a container image and talks to Postgres, it will run here. Point Clundra at the source, confirm the detected build, deploy.

Apps stop serving traffic, but your data and deployment history stay retrievable for a grace period, and a Postgres dump is one click away for as long as they do.

Build the app.
Clundra runs the rest.

Bring a repository, a container image or both. Get one secure application you can share, change and rewind.

Read the docs