Resources
Know what you are shipping. And what it runs on.
Field guides for your next deployment, long-form whitepapers, ready-to-run agent workflows, a glossary and honest comparisons. Written for the people who build the software and the people who keep it up.
- Guides
- 13 practical guides
- Whitepapers
- 3 long reads
- Workflows
- 9 agent recipes
- Glossary
- 60 terms defined
Featured whitepaper · WP-01 · 15 min read
The deploy gap.
Why software written with AI stalls before production, and what closes the gap.
- 01Writing the application is no longer the slow part. Getting it into production, and keeping it there, is.
- 02Production adds eight concerns that localhost never asks about, from reproducible builds to a named owner for the bill.
- 03Platforms-as-a-service hide those concerns until you outgrow them; do-it-yourself infrastructure asks for an SRE you do not have yet.
Guides
Practical, one topic at a time.
You built it with an AI. Here is how to ship it.
A working localhost is most of an application and very little of a deployment. What is missing, in order, and how the Kapten crew handles each step.
8 min readConnect Kapten to Claude Code, Claude Desktop and Cursor
Add the Kapten MCP server to your coding assistant, sign in, choose a scope and run your first five prompts.
7 min readBefore your first production deployment
A practical checklist for the application, the environment and the team that will operate them.
5 min readHow multi-agent workflows run
What happens between your prompt and the plan you approve, including how Kapten splits the work, how specialists run in parallel, and how to read the result.
8 min readGuardrails for agents that touch production
Scopes, approval tokens, named owners, budget caps, a scoped cloud role and an audit trail in your own account, and why plans are the defence against prompt injection.
9 min readContainerize any app: a practical Dockerfile checklist
Turn a working repository into a reproducible image. Short Dockerfiles for Node, Python and Go, and the ten rules that make them production-worthy.
8 min read
By the numbers
What is on the shelf.
Counted from the library itself when the site is built.
- 13Guides
- 3Whitepapers
- 9Agent workflows
- 60Glossary terms
- 28Integrations listed
- 138Minutes of reading
| Part | Value (guides) | Share |
|---|---|---|
| Beginner | 5 | 38% |
| Intermediate | 5 | 38% |
| Advanced | 3 | 23% |
| Total | 13 | 100% |
| Row | Via MCP | Early access | Roadmap | Not rated | Total |
|---|---|---|---|---|---|
| Agents and editors | 5 | 0 | 0 | 0 | 5 |
| Clouds | 0 | 1 | 3 | 0 | 4 |
| Source and build | 0 | 2 | 1 | 0 | 3 |
| Data and services | 0 | 4 | 0 | 2 | 6 |
| Runtimes | 0 | 7 | 0 | 0 | 7 |
| Notifications and signals | 0 | 0 | 3 | 0 | 3 |
Integrations
19 tools work today.
Whitepapers
The longer argument.
The deploy gap
Why software written with AI stalls before production, and what closes the gap.
For founders, product engineers and indie builders shipping ai-assisted codeThe agentic SRE
Operating production with a crew of specialised agents, and keeping humans in charge.
For engineering leads, platform and sre teamsOwnership by design
Bring-your-own-cloud as an architecture for SaaS teams and the customers they serve.
For ctos, saas vendors, security reviewers
Agent workflows
Recipes you can paste into your agent.
Ship your first app to production
From a repository that runs on localhost to a production environment in your own cloud account, with a domain, TLS and a rollback.
› Deploy this repo to production on my AWS account in eu-west-3. Use a managed Postgres, put the API behind api.example.com and keep it under 150 €/month.
- Navigator
- Shipwright
- Bosun
- Purser
- Warden
- Kapten
- Helmsman
Production readiness review
Four specialists review an environment in parallel and hand back one prioritised list of what to fix before you rely on it.
› Run a production readiness review on the production environment of project “atlas”. Prioritise what could lose data or take us down.
- Navigator
- Warden
- Lookout
- Purser
- Kapten
Incident triage
When something breaks, start from what changed. The crew correlates the symptom with deployments, config and dependencies, and proposes a mitigation.
› Checkout is returning 502s in production since about 14:10. What changed, and what is the safest way to stop the bleeding?
- Lookout
- Helmsman
- Kapten
Glossary
60 terms, defined plainly.
- MCP (Model Context Protocol)
- An open protocol for connecting AI assistants to tools and data.
- BYOC
- Bring your own cloud: the software runs in your cloud account.
- Multi-agent workflow
- A task split between specialised agents coordinated by an orchestrator.
- SLO
- Service level objective: a target for an SLI over a time window.
- Expand/contract migration
- Changing a database schema in backward-compatible steps.
- Prompt injection
- Instructions hidden in data that try to redirect an agent.
Comparisons
Honest trade-offs, side by side.
- Kapten vs. a traditional PaaS
- Kapten vs. DIY Terraform and CI
- Kapten vs. hiring your first SRE
- Kapten vs. a coding agent with cloud credentials
From the blog
How we think about running other people’s clouds.
BYOC is not self-hosting
The two get used interchangeably and they are opposites: one hands you the operational burden, the other hands you the account it runs in.
What a scoped role actually scopes
Every BYOC vendor says "least privilege". Here is the specific IAM shape Kapten asks for, why each permission is in it, and what it cannot reach.
The deploy should be boring
Health-gated rollouts, addressable images and rollback as a re-point rather than a re-run: the small decisions that make a deploy stop being an event.
Build with Kapten
Start with one service.
Build from there.
Tell us what you want to deploy. We review each request and follow up about fit and onboarding on your preferred cloud.
Free during early access. No card required. Cloud usage billed separately.