ChiselStrike is a complete backend bundled in one piece. Your one stop-shop for all your backend needs, powered by TypeScript.
Putting together a backend is hard work. Databases? ORM? Business logic? Data access policies? And how to offer all of that through an API?
Learning all that, plus figuring out the interactions between all the components can be a drain on an application developer's time. Low-code based approaches allow for super fast prototypes, but as you need to scale and evolve, the time you saved on the prototype is now owed with interest in refactorings, migrations, etc.
ChiselStrike provides everything you need to handle and evolve your backend, from the data layer to the business logic, allowing you to focus on what you care about – your code, rather than worrying about databases schemas, migrations, or even database operations.
All driven by TypeScript, so your backend can evolve as your code evolves.
ChiselStrike keeps things as close as possible to pure TypeScript, and a translation layer takes care of index creation, database query generation, and even communicating with external systems like Kafka.
Internally, ChiselStrike uses a SQLite database so there's no need to set up any external data layer (although it is possible to hook up an external Postgres-compatible database). ChiselStrike also abstract other concepts common to complex backends, like Kafka-compatible streaming platforms.
To get a CRUD API working in 30 seconds or less, first create a new project:
npx -y create-chiselstrike-app@latest my-app
cd my-appAdd a model by writing the following TypeScript code to models/BlogComment.ts:
import { ChiselEntity } from "@chiselstrike/api"
export class BlogComment extends ChiselEntity {
content: string = "";
by: string = "";
}Add a route by writing the following TypeScript code to routes/comments.ts:
import { BlogComment } from "../models/BlogComment";
export default BlogComment.crud();Start the development server with:
npm run devThis server will provide a CRUD API that you can use to add and query instances of the BlogComment entity.
curl -X POST -d '{"content": "First comment", "by": "Jill"}' localhost:8080/dev/comments
curl localhost:8080/dev/commentsFor a more detailed tutorial about how to get started with ChiselStrike, follow our Getting started tutorial.
No. The founding team at ChiselStrike have written databases from scratch before and we believe there are better things to do in life, like pretty much anything else. ChiselStrike comes bundled with SQLite, providing developers with a zero-conf relational-like abstraction that allows one to think of backends from the business needs down, instead of from the database up.
Instead, you can think of ChiselStrike as a big pool of global shared memory. The data access API is an integral part of ChiselStrike and offers developers a way to just code, without worrying about the underlying database (anymore than you worry about what happens in each level of the memory hierarchy, meaning some people do, but most people don't have to!).
In production, ChiselStrike can also hook up into a Kafka-compatible streaming platform when available, and transparently drive both that and the database from a unified TypeScript/JavaScript abstraction.
Kind of. ChiselStrike has some aspects that overlap with traditional ORMs, in that it allows you to access database abstractions in your programming language. However, in traditional ORMs you start from the database, and export it up. Changes are done to the database schema, which is then bubbled up through migrations, and elements of the database invariably leak to the API.
ChiselStrike, on the other hand, starts from your code and automates the decisions needed to implement that into the database, much like what a compiler would do.
Let's look at ChiselStrike's documentation for an example of what's needed to create a comment on a blog post:
import { ChiselEntity } from "@chiselstrike/api"
export class BlogComment extends ChiselEntity {
content: string = "";
by: string = "";
}The first thing you will notice is that there is no need to specify how those things map to the underlying database. No tracking of primary keys, column types, etc.
Now imagine you need to start tracking whether this was created by a human or a bot. You can change your model to say:
import { ChiselEntity } from "@chiselstrike/api"
export class BlogComment extends ChiselEntity {
content: string = "";
by: string = "";
isHuman: boolean = false;
}and that's it! There are no migrations and no need to alter a table.
Furthermore, if you need to find all blog posts written by humans, you can just write a lambda instead of trying to craft a database query in TypeScript:
const all = await BlogComment.findMany(p => p.isHuman);ChiselStrike includes a TypeScript runtime - the fantastic and beloved Deno. That's the last piece of the puzzle with the data API and the database bundles. That allows you to develop everything locally from your laptop and integrate with your favorite frontend framework. Be it Next.js, Gatsby, Remix, or any others - we're cool with all of them!
We hear you. No modern application is complete without authentication and security. ChiselStrike integrates with next-auth and allows you to specify authentication entities directly from your TypeScript models.
You can then add a policy file that details which fields can be accessed, and which endpoints are available.
For example, you can store the blog authors as part of the models,
import { ChiselEntity, AuthUser } from "@chiselstrike/api"
export class BlogComment extends ChiselEntity {
content: string = "";
@labels("protect") author: AuthUser;
}and then write a policy saying that the users should only be able to see the posts that they themselves originated:
labels:
- name: protect
transform: match_loginNow your security policies are declaratively applied separately from the code, and you can easily grasp what's going on.
ChiselStrike provides everything you need to handle your backend, from the data layer to the business logic, wrapped in powerful abstractions that let you just code and not worry about handling databases schemas, migrations, and operations again.
It allows you to declaratively specify compliance policies around who can access the data and under which circumstances.
Your ChiselStrike files can go into their own repo, or even better, into a subdirectory of your existing frontend repo. You can code your presentation and data layer together, and turn any frontend framework into a full-stack (including the database layer!) framework in minutes.
To build and develop from source:
git submodule update --init --recursive
cargo buildThat will build the chiseld server and chisel utility.
You can now use create-chiselstrike-app to install a local version of the API:
node ./packages/create-chiselstrike-app --chisel-version="file:../packages/chiselstrike-api" my-backendAnd then replace instances of npm run with direct calls to the new binaries. For example, instead of
npm run dev, run
cd my-backend
npm i esbuild
../target/debug/chisel devAlso, consider:
Start a discussion 🙋♀️
Our documentation (including a quick tutorial) is available at docs.chiselstrike.com
Don't want the hassle of deploying and managing this yourself?
Our hosted version is available here and is now in beta. You can deploy your project straight from your git repository and see your backend materializing in front of your eyes.