Capsules
A production for every branch.
Zero-config preview environments on your cloud. Every git branch gets a real deployment — same services, same networks, same TLS — at a unique URL.
What It Is
Production minus the users.
Same cloud, same topology, same container images, same secrets, same deployment pipeline. The only difference is the branch.
Not a namespace. Not a container. Not a Docker Compose approximation. A real, running deployment on the infrastructure your production already uses — scoped to a single branch.
The Status Quo
Your branch doesn't have its own cloud.
Production has one. Staging has one, sort of. Everything else lives in a spreadsheet of ad-hoc environments and hope.
The Staging Queue
Someone's testing a migration; your PR waits. Nothing ships until one person finishes.
Works on My Compose
OAuth callbacks, CORS, security groups, cert expiry — none of it shows up locally, all of it shows up in prod.
The $900 Ghost
The forgotten RDS instance nobody remembered to tear down, running up the monthly bill in silence.
The Unclickable PR
Reviewed as a green check and a diff, because there was nowhere to actually click the feature.
How It Works
Four steps. Then git.
01
Open your project
Monk reads your code — framework, services, data stores, integrations — and writes a manifest you check in.
02
Set up Capsules
Say `set up Capsules` in Monk chat. Pick a cloud and region. Monk generates the workflow and commits it.
03
Push a branch
A few minutes later, a preview URL shows up on the PR. Real deployment, real infrastructure.
04
Merge or delete
The capsule tears itself down. Secrets revoked, resources released. Nothing lingers.
Impact
What changes the day you turn them on.
PRs are clickable.
The proof is a working link, not a green checkmark. Reviewers, product, QA open the URL and see the thing.
Environment drift disappears.
Staging for this branch is shaped like prod — same managed services, same networks, same TLS. The bugs surface early.
Ten branches, ten productions.
No queueing, no conflicts. Parallel work becomes actually parallel — each branch has its own cloud.
Agents verify their own work.
Push, hit real endpoints, read real logs, fix, push again. The PR an agent hands off is tested, not guessed.
Infra changes become PRs.
Swap the cache engine, try a smaller VM, benchmark a database. The capsule rebuilds. Merge when the preview works.
Nothing lingers.
Branch deleted → capsule gone. No forgotten RDS. No end-of-quarter cleanup sprint.
Why It Works On Day One
Not a wrapper over terraform apply.
Three things have to be true at once. Monk gets them all right — which is why capsules work out of the box instead of after a platform team's quarter.
01
Typed, lifecycle-aware primitives
Every cloud resource is a primitive that knows how to provision, configure, connect, and tear itself down. The graph for your app builds itself from your code.
02
A persistent orchestrator with state
Capsules remember what they created. Drift and lifecycle are reasoned about over time — not a shell script, not an LLM conversation.
03
Scoped credentials
A branch never holds your full production secret bundle. Per-capsule, per-entity, short-lived tokens minted at runtime.
This is why capsules run on AWS, Azure, GCP, and DigitalOcean on the same day. The orchestrator is the portability.
Practical
Additive. Ephemeral. Yours.
Cost control built in
Hibernate on a schedule — weekdays only, off overnight. Override per capsule when a demo runs late.
GitHub-native
Each capsule registers as a GitHub Deployment. Secrets live in encrypted environment secrets. Nothing new to learn.
Doesn't touch your setup
Branches you want left alone — main, release branches — are excluded by filter. Turn capsules off tomorrow; production is exactly where you left it.
A New Category
Autonomous software.
Software that builds, deploys, and manages itself. No human in the execution loop. You define the intent. You set the guardrails. Your coding agent and Monk handle everything between idea and production.
5-day free trial.