Define collections, fields, access rules and hooks in Go. Ridu generates an admin, a REST API, database migrations and TypeScript types from that config, and builds the whole thing into one binary.
npm create ridu@latest
How it works
var Posts = ridu.Collection{
Slug: "posts",
Fields: field.Fields{
field.Text("title").Required(),
field.Textarea("summary").MaxLength(280),
field.Select("status", "draft", "published"),
},
}Generated from this file
const page = await ridu.list('posts', {
where: { status: { equals: 'published' } }
});{
"docs": [
{
"id": "posts_7514213d859111e0",
"title": "Hello from Ridu",
"status": "published"
}
],
"totalDocs": 1
}$ npm run build
[ridu] Verified migration artifact history
Built dist/app
$ ls -lh dist/app
-rwxr-xr-x 1 you staff 28M dist/appBehaviour in code
They run for every request, whether it comes from the admin, the REST API, the SDK or your own Go code. Here are four small examples and what each one does.
Signed-out visitors only get published posts. The filter is applied in the database query.
Read: func(ridu.AccessContext) (
ridu.AccessDecision, error,
) {
return ridu.Where(
query.Equal("status", "published"),
), nil
},What a signed-out visitor gets
Drafts aren't returned, and they aren't included in the total count.
Trims the title before it's saved, whichever API the save came from.
func(ctx ridu.HookContext) error {
title, _ := ctx.Data["title"].StringValue()
ctx.Data["title"] = store.String(
strings.TrimSpace(title),
)
return nil
}Saved title
Add a value to API responses without storing it.
field.Virtual("displayName",
func(ctx operation.Context) (
operation.Value[string], error,
) {
first, _ := ctx.Root.String("firstName")
last, _ := ctx.Root.String("lastName")
name := first + " " + last
return operation.Present(name), nil
})GET /api/people/ada
{
"firstName": "Ada",
"lastName": "Lovelace",
"displayName": "Ada Lovelace"
}Define the blocks editors can add to a page, and the fields in each one.
field.Blocks("layout",
field.Block{Slug: "hero", Fields: heroFields},
field.Block{Slug: "quote", Fields: quoteFields},
field.Block{Slug: "gallery", Fields: galleryFields},
)In the admin
Deploy
ridu build produces one executable with the API, the admin and background jobs in it. Nothing else runs in production, so the image only needs that file. Migrations are generated from your config, and the binary checks they have been applied before it starts. There's also aRailway guide.
Database
FROM gcr.io/distroless/static-debian12A static base image with no shellCOPY dist/my-app /appYour app, with the admin built inENV DATABASE_URL=postgres://…ENV RIDU_SQLITE_PATH=/data/app.sqliteENV DATABASE_URL=mongodb+srv://…Your PostgreSQL connection URLA database file on a mounted volumeNeeds a replica set and TLSEXPOSE 8080Health checks use /readyzENTRYPOINT ["/app"]Checks migrations are applied firstComing from Payload?
Collections, fields, access rules, hooks, drafts, versions and live preview work the way you'd expect. The migration guide goes through each Payload concept and shows the Ridu version next to it.
fields: [
{ name: 'title', type: 'text', required: true },
{ name: 'slug', type: 'text', unique: true },
{ name: 'summary', type: 'textarea' },
{ name: 'author', type: 'relationship',
relationTo: 'users' },
],Fields: field.Fields{
field.Text("title").Required(),
field.Text("slug").Unique(),
field.Textarea("summary"),
field.Relationship("author", "users"),
},Pick the starter template and SQLite and you won't need Docker or a database server. The quickstart walks through creating a post in the admin and reading it back with the generated SDK.
npm create ridu@latest my-app