Ginboot GitHub auto build
Install the GitHub App, pick a repo and branch, and every push is built automatically. Build logs stream live, and you can retry or cancel a build with one click.
Serverless deploys for Go and frontends
Push to GitHub.
Push to GitHub.
Run on Lambda.
Ginboot Cloud builds every commit, deploys your API to AWS Lambda and your frontend to CloudFront, and gives you the metrics, traces and alerts to run them. You don't need to write YAML or manage a pipeline. You can use your own AWS account if you want.
~/orders-api git push origin main→ github.com/acme/orders-api a3f9c21..7be04d1 ginboot build #184 started from 7be04d1 ▸ go build ./cmd/api ✓ 18.2s ▸ vite build ./web ✓ 11.7s ▸ artifact sealed sha256:9c1e… ✓ ginboot deploy #184 → preprod λ orders-api Lambda · 512 MB live ☁ web CloudFront live ginboot promote #184 preprod → prod ✓ same artifact, zero rebuild · 4.1s https://api.acme.dev
Runs on
AWS LambdaAPI GatewayCloudFrontOpenTelemetryGitHubMCP
Platform
Everything between git push and production
Ginboot Cloud handles building, deploying, monitoring and access control, so your team can spend its time on the application.
Lambda-first design
APIs run on AWS Lambda behind API Gateway. They scale from zero, you pay nothing while they sit idle, and there are no servers to patch.
Frontend and backend
Deploy a Go API and a static frontend in the same project. Frontends are served from CloudFront's global edge network, and you can attach custom domains.
Built-in telemetry
See requests, error rates, latency and compute for every app without adding an SDK. The numbers come straight from API Gateway, Lambda and CloudFront.
Error reporting
Set an alert per app and environment, such as "5 errors in 10 minutes", and the right people get an email. Returned 500s count as errors, not only crashes.
Debugging and traceability
Follow a request across spans, jump from a trace to its logs, and trace any deployment back to the commit and build that produced it.
AI agent support for your favourite IDE
A remote MCP server gives Claude Code, Cursor, VS Code and Windsurf the same view of your platform that you have. Your agent can read logs and traces, start builds and promote releases. Changes are limited by your role and need your approval.
ginboot_query_tracesginboot_get_build_logsginboot_promote_buildginboot_rollbackPromote the same build
Create as many environments as you need. Each build is immutable, so the artifact you tested in preprod is exactly what ships to prod.
Connect your own AWS
Connect your AWS account with one CloudFormation template. Your apps then run in your account, under your limits and on your AWS bill.
Serverless databases
Create PostgreSQL-compatible Aurora DSQL databases in your connected account, then browse tables, edit rows and run queries from the console. Read the docs
Permission models
Build custom roles from fine-grained permissions such as app:deploy and billing:read, and invite teammates with exactly the access they need. In connected accounts, you set each app's AWS access per environment in the console, and a permission boundary caps it.
Workflow
Build once, promote everywhere
Other platforms rebuild for each environment. Ginboot builds your code once and moves that same artifact through every environment, so production runs exactly what you tested.
01 · push
Commit to GitHub
Push to a tracked branch to start a build, or trigger one manually from the console or your IDE.
7be04d102 · build
Seal an artifact
The API and frontend are compiled, packaged and versioned as build #184.
build #18403 · verify
Deploy to preprod
It runs with its own variables, secrets and domain. Check it with Pulse and traces before you go further.
preprod04 · ship
Promote to prod
The same artifact moves to prod without a rebuild. One click rolls it back if something looks wrong.
prod · #184// environments are just names: dev, qa, staging, eu-prod… add as many as you need
Observability
Find out what broke, and which commit broke it
You get metrics, traces, logs and alerts on the day you deploy, with nothing to set up.
- Pulse: requests, 4xx/5xx, p95 latency and GB-seconds for every app, read from the metrics AWS already records.
- Traces and logs: OpenTelemetry is built in. Open a slow span and see the logs for that request.
- Route stats: find the endpoint that is slowing things down or returning errors.
- Error alerts: per-environment thresholds, emailed to the people who own the app.
AI agents
Your coding agent can see production too
Add Ginboot's MCP server to your editor. Your agent can then find a failing request, read its trace, fix the code, ship a build and confirm the fix, all from your IDE.
- Works with any MCP client: Claude Code, Cursor, VS Code, Windsurf and others.
- Reading is instant: projects, builds, logs, traces and route stats.
- Guarded changes: before a deploy, promotion or rollback, the agent must name the exact build and environment, and your editor asks you to approve. Protected environments are refused outright, and your role's permissions always apply.
- OAuth sign-in: no long-lived API keys in config files.
$ claude mcp add --transport http ginboot https://api.ginboot.cloud/mcp
Access control
Give each person the access they need
Roles are made from fine-grained permissions, so a contractor can deploy to preprod without being able to see billing or invite users.
- Custom roles: combine granular permissions into whatever roles your team needs.
- Invites: add teammates by email and assign a role at the same time.
- App AWS permissions: in a connected account, grant an app exactly the AWS access it needs for each environment. A permission boundary caps it.
Bring your own AWS
Run your apps in your own AWS account
Start on our account. When you need your own billing, limits or compliance boundary, deploy one CloudFormation template and connect your account. Environments that are already running stay where they are.
- CodeBuild — compiles your commit
- Artifact store — sealed builds
- Pulse, alerts, traces
assume role
scoped · revocable
scoped · revocable
- Lambda + API Gateway
- S3 + CloudFront
- Your quotas, your AWS bill
Pricing
Pay for usage, not seats
Every organization gets a generous monthly allowance, and there's no charge per team member. Above the allowance, you pay close to AWS cost.
Hosted
We run everything in Ginboot's AWS account. The fastest way to ship.
$0/ month + usage
- 1M API requests and 100k GB-s compute included
- 10M CDN requests and 100 GB transfer included
- 300 build minutes / month
- Unlimited projects, environments and teammates
- Telemetry, traces, error alerts and MCP
Your AWS
Your apps run in your account and AWS bills you directly for runtime. We operate the platform.
$49/ month per organization
- Everything in Hosted
- No Ginboot charges on requests, compute or CDN
- 300 build minutes included, then $0.008/min
- Your quotas, VPC limits and compliance boundary
- Serverless databases, billed by AWS
- Scoped, revocable cross-account role
| Meter | Included / mo | Then |
|---|---|---|
| API requestsAPI Gateway | 1,000,000 | $1.50 / million |
| ComputeLambda duration × memory | 100,000 GB-s | $0.000025 / GB-s |
| CDN requestsCloudFront | 10,000,000 | $1.20 / million |
| CDN transferCloudFront egress | 100 GB | $0.10 / GB |
| Build minutesAll plans | 300 min | $0.008 / min |
Estimate your API bill
Hosted plan, backend only. Move the sliders.
Estimated / month$7.25
Prices in USD, metered per calendar month (UTC). Rates are introductory and may change with notice.
FAQ
Common questions
What can I deploy on Ginboot Cloud?
Go APIs built with the Ginboot framework, which run on AWS Lambda behind API Gateway, and static frontends such as React, Vue or Svelte, which are served from CloudFront. Both are built from your GitHub repository. To start a new app, use the Ginboot Starter.
Do I need an AWS account?
No. By default your apps run in Ginboot's AWS account. You can connect your own account at any time with a CloudFormation template. New environments then deploy there and AWS bills you directly.
Which AI coding tools work with Ginboot Cloud?
Any tool that supports the Model Context Protocol (MCP), including Claude Code, Cursor, VS Code and Windsurf. Read-only tools work immediately. Builds, deploys, promotions and rollbacks wait for your approval.
Do I need to instrument my app to get metrics?
No. Request, error, latency and compute metrics come from what API Gateway, Lambda and CloudFront already record. Apps built with Ginboot also export OpenTelemetry logs and traces for request-level debugging.
Can Ginboot create a database for my app?
Yes, if you've connected your own AWS account. Ginboot creates an Amazon Aurora DSQL cluster in that account. It's a serverless, PostgreSQL-compatible database that AWS bills to you, and it costs only its storage while idle. You can connect any Postgres client to it, and browse, edit and query its data from the console. The databases docs cover connecting, billing, limits and moving an existing Postgres database.
How does promoting a build work?
Each push produces one immutable build. You deploy it to preprod, check it, and then promote that same artifact to production. What you tested is byte-for-byte what ships, and rolling back means redeploying the previous build.
Ship your next commit in a few minutes
Connect GitHub, pick a repo and push. Your first build can be live on Lambda the same afternoon.