fix(deps): update prisma monorepo to v6 (main) (major)

[renovate]

This PR contains the following updates:

Package	Change	Age	Confidence
@prisma/client (source)	4.16.2 -> 6.18.0
prisma (source)	4.16.2 -> 6.18.0

---

Release Notes

prisma/prisma (@​prisma/client)

v6.18.0

Compare Source

Today, we are excited to share the 6.18.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Prisma ORM

Prisma ORM is the most popular ORM in the TypeScript ecosystem. Today’s release brings a bunch of new bug fixes and overall improvements:

• prisma init now creates a prisma.config.ts automatically

When creating a new project with 6.18.0, prisma init will now create a prisma.config.ts file automatically. This prepares new applications for the future of Prisma 7. Some fields that have been historically set in the schema.prisma file are now able to be set in the prisma.config.ts, and we encourage people to migrate over to the new structure before the release of version 7, where this file will become a requirement.

• Support for defining your datasource in prisma.config.ts

If you’re adopting the new prisma.config.ts setup in your projects, version 6.18.0 brings the ability to set your datasource directly in your config file. Once this is in your config file, any datasource set in your schema.prisma will be ignored. To set the datasource, we also must include the new engine key which we can set to "classic" , which will be required for Prisma v7

import { defineConfig, env } from "prisma/config";
export default defineConfig({
    // The Rust-compiled schema engine 
    engine: "classic",
    datasource: {
        url: env('DATABASE_URL'),
    }
});

• #​28291 Support multiple Prisma instances with different providers
• #​28305 Add env helper function
• #​28266 Add support for js or classic as engine types in prisma.config
• #​28139 Map Bytes to Uint8Array depending on Typescript version

Preparing for Prisma v7

While it has been mentioned a few times already, many of the changes in this release are here to prepare folks for the upcoming release of Prisma v7. It’s worth repeating that these changes and the migration to prisma.config.ts will be required for Prisma v7, so we’re releasing this as opt-in features for developers. But come Prisma v7, they will be the new way of configuring your project.

Prisma Postgres

Prisma Postgres is our fully managed Postgres service designed with the same philosophy of great DX that has guided Prisma for close to a decade. With this release we are introducing the following improvements:

Database Metric in Console

Inside of your database console, you can now view metrics on your database usage and interactions. You can get insights into the follow:

• Total egress
• Average response size
• Average query duration

In addition, you can also get insights into how to improve your query caching and gain better performance.

Open roles at Prisma

Interested in joining Prisma? We’re growing and have several exciting opportunities across the company for developers who are passionate about building with Prisma. Explore our open positions on our Careers page and find the role that’s right for you.

Enterprise support

Thousands of teams use Prisma and many of them already tap into our Enterprise & Agency Support Program for hands-on help with everything from schema integrations and performance tuning to security and compliance.

With this program you also get priority issue triage and bug fixes, expert scalability advice, and custom training so that your Prisma-powered apps stay rock-solid at any scale. Learn more or join: https://prisma.io/enterprise.

v6.17.1

Compare Source

Today, we are issuing a patch release to address a regression in v6.17.0 that affected diffing of unsupported types, leading to unnecessary or incorrect changes when creating new migrations or running db pull. This update is recommended for all users who have any fields marked as Unsupported in their schema files.

Changes

• prisma/prisma-engines#5633

v6.17.0

Compare Source

Today, we are excited to share the 6.17.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Prisma ORM

Prisma ORM is the most popular ORM in the TypeScript ecosystem. Today's release brings a number of bug fixes and improvements to Prisma ORM.

Bug fixes and improvements

• Added support for Entra ID (ActiveDirectory) authentication parameters for the MS SQL Server driver adapter. For example, you can use the config object to configure DefaultAzureCredential:

  import { PrismaMssql } from '@​prisma/adapter-mssql'
  import { PrismaClient } from '@​prisma/client'
  
  const config = {
    server: 'localhost',
    port: 1433,
    database: 'mydb',
    authentication: {
      type: 'azure-active-directory-default',
    },
    options: {
      encrypt: true,
    },
  }
  
  const adapter = new PrismaMssql(config)
  const prisma = new PrismaClient({ adapter })

  Learn more in this PR.
• Relaxed the support package range for @opentelemetry/instrumentation to be compatible with ">=0.52.0 <1". Learn more in this PR.
• Added Codex CLI detection, ensuring dangerous Prisma operations are not executed by Codex without explicit user consent. Learn more in this PR.
• Fixed JSON column handling when using a MariaDB database. Learn more in this PR.
• Restored the original behaviour of group-by aggregations where they would refer to columns with explicit table names which fixes a regression that would result in ambiguous column errors. Learn more in this PR.

Prisma Postgres

Prisma Postgres is our fully managed Postgres service designed with the same philosophy of great DX that has guided Prisma for close to a decade. With this release we are introducing the following improvements:

New usage workspace metrics available in your Console Dashboard

The Dashboard in your Prisma Console account now displays new metrics about your Prisma Postgres usage:

• Key metrics
  • Estimated upcoming invoice
  • Total storage used
  • Total DBs
• Overall usage
  • Cumulative operations
  • Operations per day

Using Prisma Postgres with any tool is ready for production

Previously, the only way to connect to Prisma Postgres was using Prisma ORM. That combination is great because it gives you connection pooling, global caching and overall an amazing DX.

That being said, we understand that preferences vary and some developers prefer to use plain SQL or lower-level query builders in their applications. As of this release, these ways for connecting to Prisma Postgres are now officially generally available and can be used in your production apps!

You can connect using Drizzle, Kysely, TypeORM, psql, or any other Postgres-compatible library, database migration tools like Atlas or interfaces like DBeaver, Postico, and more.

📚 Learn more in the docs.

Enterprise support

Thousands of teams use Prisma and many of them already tap into our Enterprise & Agency Support Program for hands-on help with everything from schema integrations and performance tuning to security and compliance.

With this program you also get priority issue triage and bug fixes, expert scalability advice, and custom training so that your Prisma-powered apps stay rock-solid at any scale. Learn more or join: https://prisma.io/enterprise.

v6.16.3

Compare Source

Today, we are issuing a 6.16.3 patch release focused on bug fixes.

🛠 Fixes

• Prisma Client (prisma-client generator): fixed missing JSON null type definitions (JsonNull, DbNull, AnyNull) in the browser.ts entrypoint. (#​28186)

• Prisma Migrate: don't add the default schema (namespace) to the generated migrations unless it was specified explicitly in the schema file. This restores the pre-6.13.0 behaviour that was inadvertently changed with enabling multi-schema support by default. Users who rely on database schemas for multi-tenancy can now again use the same migration files for all of their schemas. (prisma/prisma-engines#5614)

• Prisma Client: enabled negative take with findFirst again. (prisma/prisma-engines#5616 — contributed by @​jay-l-e-e)

• Prisma Accelerate: aligned the behaviour of the new Rust-free client with Query Engine to handle self-signed certificates consistently and ensure backward compatibility. (#​28134)

• @prisma/adapter-mariadb: fixed error event listeners leak. (#​28177 — contributed by @​Tiaansu)

⚠️ Known Limitation: JSON null types in browser builds

The fix introduces the missing types, but the singleton instances differ between the client and browser entrypoints of the generated client. This means that values like Prisma.JsonNull imported from browser cannot yet be assigned to fields expected from the client entrypoint, and vice versa. This results in confusing TypeScript errors if you mix them. A follow-up improvement is planned to unify these utility types across entrypoints.

v6.16.2

Compare Source

Today, we are issuing a 6.16.2 patch release.

Bug fixes

• In Prisma ORM 6.16.0, we've enabled usage of the new engineType = client with Prisma Postgres, but our validation rules permitted invalid combinations of Prisma Postgres URLs and driver adapters. This now produces a clear error message indicating Prisma Postgres URLs and driver adapters are mutually exclusive.
• In the previous minor release, we've included a change that calls unref() on NodeJS timers to prevent them from keeping the NodeJS event loop active. This change unintentionally affected non-NodeJS runtimes like workerd, where it has resulted in runtime errors. This behavior has been made conditional to prevent these runtime errors.

v6.16.1

Compare Source

Today, we are issuing a 6.16.1 patch release.

Bug fixes

• In Prisma ORM 6.16.0, the driverAdapters and queryCompiler features were stabilized, but leftover code in the prisma-client-ts generator required them to still be specified in edge runtimes. This has now been fixed, runtimes like workerd and vercel-edge no longer require these preview features.

v6.16.0

Compare Source

Today, we are excited to share the 6.16.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Prisma ORM

This section contains all the updates made in Prisma ORM v6.16.0.

Rust-free ORM and driver adapters are Generally Available

Eight months ago, we published our ORM manifesto with the first hint that we're going to remove the Rust-based query engine from Prisma ORM:

We're addressing this by migrating Prisma's core logic from Rust to TypeScript and redesigning the ORM to make customization and extension easier.

After a lot of hard work and feedback from the community, we're incredibly excited to share that the migration has been completed and you can now use Prisma ORM without its Rust engine in your production apps. 🎉 This is a major milestone in the history of Prisma ORM and comes with a lot of benefits:

• Reduced bundle size by ~90%
• Faster queries (check out our latest benchmarks)
• Lower CPU footprint
• Less deployment complexity
• Easier to make open-source contributions

… and overall a much better DX since you don't need to worry about the extra binary in your generated Prisma Client code any more.

While the Rust-free ORM will become the default in Prisma ORM v7 soon, for now you still need to opt-into using it:

1. Configure the generator block in your Prisma schema:

   generator client {
     provider   = "prisma-client" // (or "prisma-client-js") 
     output     = "../src/generated/prisma"
     engineType = "client"
   }

   Note: If you tried the Rust-free ORM before, you can now also drop the queryCompiler and driverAdapter feature flags from the previewFeatures array. And if you used binaryTargets, you can also get rid of these.
2. Install the driver adapter for your database, e.g. to use pg for PostgreSQL:

   npm install @​prisma/adapter-pg
   

3. Finally, you can instantiate PrismaClient using the PrismaPg driver adapter as follows:

   import { PrismaClient } from './generated/prisma'
   import { PrismaPg } from '@​prisma/adapter-pg'
   
   const adapter = new PrismaPg({ connectionString: env.DATABASE_URL })
   const prisma = new PrismaClient({ adapter })
   
   // ... send queries using `prisma` like before

📚 To learn more and see instructions for all other supported databases, check out the documentation.

The Rust-free version of Prisma ORM has been thoroughly tested with the prisma-client generator (see below), not with prisma-client-js. Use the old generator at your discretion.

New ESM-first prisma-client generator is Generally Available

Another major milestone has been achieved in this release: The new, flexible and ESM-first prisma-client generator is ready for production too. Here's a quick overview of its main benefits:

• No more magic generation into node_modules; generated code is fully under control by the developer
• ESM-compatible by default
• Flexible configuration for specific runtimes (Node.js, Deno, Bun, Cloudflare, …)

generator client {
  // Required
  provider = "prisma-client"
  output   = "../src/generated/prisma"

  // Optional
  engineType             = "client"
  runtime                = "nodejs"
  moduleFormat           = "esm"
  generatedFileExtension = "ts"
  importFileExtension    = "ts"
}

In addition to making it production-ready, we also made some changes to the prisma-client generator:

• removed Prisma.validator; you can use TypeScript native satisfies keyword instead
• created a new ./generared/prisma/browser entrypoint for importing types in browser environments

If you want to try out the new generator with your favorite framework, check out one of our ready-to-run examples (e.g. for Next.js, Nuxt or React Router).

📚 Learn more in the docs.

Type check performance optimizations

Runtime performance is not the only performance category that matters. In fact, when it comes to DX, type checking performance is equally important: if your TypeScript types become too complex and the compiler needs to do too much work (e.g. inferring types), it may slow down your editor, lead to laggy auto-completion or prevent jump-to-definition from working.

We've worked with TypeScript expert David Blass to find ways for improving the type checking performance in Prisma ORM and created benchmarks comparing the type checking performance with Drizzle.

You can read about the results here: Why Prisma ORM Checks Types Faster Than Drizzle

Deprecating the postgresqlExtensions Preview feature

We're deprecating the postgresqlExtensions Preview feature. Note that this doesn't mean that you can't use extensions with Prisma ORM any more. Instead of setting the Preview feature, you can install extensions manually with a customized migration via the --create-only flag:

npx prisma migrate dev --name add-extension --create-only

You can then install an extension with plain SQL in the newly created, empty migration file:

CREATE EXTENSION IF NOT EXISTS "pgcrypto";

Prisma Postgres

Prisma Postgres is our fully managed Postgres service designed with the same philosophy of great DX that has guided Prisma for close to a decade. With this release we are introducing the following improvements:

Manage OAuth apps in Prisma Console

In Prisma Console, you can now manage all of the 3rd party applications that you've granted access to perform actions on behalf of yourself in your Prisma Console account. Find the 🧩 Integrations tab in the sidenav to see which applications currently have access.

Rust-free Prisma ORM with Prisma Accelerate and Prisma Postgres

With this release, the Rust-free Prisma ORM (Query Compiler) can now be used together with Prisma Postgres and also Prisma Accelerate. This means you can take advantage of connection pooling and caching while using the new TypeScript-based ORM architecture.

To enable it, update your Prisma schema:

generator client {
  provider   = "prisma-client"
  output     = "../src/generated/prisma"
  engineType = "client"
}

We'd love for you to try this out and share your feedback as we prepare for General Availability. Please open an issue on GitHub if you encounter any problems or have suggestions.

Enterprise support

Thousands of teams use Prisma and many of them already tap into our Enterprise & Agency Support Program for hands-on help with everything from schema integrations and performance tuning to security and compliance.

With this program you also get priority issue triage and bug fixes, expert scalability advice, and custom training so that your Prisma-powered apps stay rock-solid at any scale. Learn more or join: https://prisma.io/enterprise.

v6.15.0

Compare Source

Today, we are excited to share the 6.15.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Highlights

AI safety guardrails for destructive commands

Prisma ORM now includes built-in safety checks that protect against destructive commands when triggered by AI coding assistants. The CLI can recognize when it is being executed by popular AI agents such as Claude Code, Gemini CLI, Qwen Code, Cursor, Aider and Replit.

If a command like prisma migrate reset --force is attempted, Prisma ORM will prompt for explicit confirmation before proceeding.

This feature ensures that irreversible operations which drop and recreate the database are not executed automatically by an AI tool. Prisma ORM is the first ORM to provide this level of protection, making it safer to use AI-assisted development while working with your databases.

📚 Learn more in the docs.

prisma-client: runtime improvements and schema flexibility

We simplified Prisma ORM by making the runtime options for the Prisma Client more consistent and easier to understand. Previously there were several overlapping aliases which created confusion. With this release we simplified the inputs while keeping support for all the major environments you might be targeting.

Changes include:

• node has been removed, use runtime = "nodejs" instead
• deno-deploy has been removed, use runtime = "deno" instead
• vercel has been replaced by the new runtime = "vercel-edge"
• edge-light is now just an alias for vercel-edge
• nodejs, deno, and bun now share the same internal code path, while still keeping their separate input values for clarity
• The VS Code extension has been updated to reflect these changes

The updated list of supported runtimes is now:

nodejs, deno, bun, workerd (alias cloudflare), vercel-edge (alias edge-light), and react-native.

In addition, we fixed an issue where running prisma generate would fail if your schema contained no models. This is now supported with the new prisma-client generator, just like it already worked with the older prisma-client-js generator.

For example, the following schema will now generate a client without errors:

generator client {
  provider = "prisma-client"
  output   = "../generated/client"
}

datasource db {
  provider = "postgresql"
  url      = env("DATABASE_URL")
}

Running prisma generate with this schema will succeed and create the client in ./generated/client.

📚 Learn more in the docs.

Using Prisma ORM with Vercel Fluid

Fluid compute is a new compute model from Vercel that combines the flexibility of serverless with the stability of servers, making it ideal for dynamic workloads such as streaming data and AI APIs.

A common challenge in traditional serverless platforms is that when functions are suspended, database connection pools can’t close idle connections. This leads to leaked connections that stay open until the database times them out, which can exhaust the pool.

Vercel provides the attachDatabasePool utility to solve this problem. It ensures idle connections in the pool are properly released before a function is suspended, preventing connection leaks.

You can use this utility together with Prisma’s driver adapters to safely manage database connections in Fluid Compute:

import { Pool } from "pg";
import { attachDatabasePool } from "@​vercel/functions";
import { PrismaPg } from "@​prisma/adapter-pg";
import { PrismaClient } from "./generated/prisma/client";

const pool = new Pool({ connectionString: process.env.POSTGRES_URL });
attachDatabasePool(pool);

const prisma = new PrismaClient({
  adapter: new PrismaPg(pool),
});

📚 Learn more in the docs.

Other news

Prisma Postgres Management API is Generally Available

The Prisma Postgres Management API allows you to programmatically provision and manage Prisma Postgres instances. It’s the perfect way to spin up a database in your CI/CD workflow, see our GitHub Action examples for creating and deleting if you’re curious about this use case.

It also enables developers to offer databases to their own users! For example, did you know that Co.dev (YC23), a popular “low-code AI app builder” is using the Management API to provision Prisma Postgres instances to people building apps with their platform?

We’re excited to share that the Management API is now fully ready for production. With it moving into GA, we also added another piece of functionality where you can now create new projects without a default database.

We’re looking forward to see what you’re going to build with it!

📚 Learn more in the docs.

Prisma Postgres is now available on Pipedream

Prisma Postgres can now be used directly in your Pipedream workflows 🎉

With this integration, you can connect Prisma Postgres to over 2,800+ apps supported on Pipedream, enabling powerful automations and data workflows. For example, you can:

• Automatically spin up a new Prisma Postgres database when a customer signs up in Stripe.
• Connect Prisma Postgres with Slack, Notion, Airtable, or any other app in the Pipedream ecosystem

This makes it easier than ever to use Prisma Postgres in your automation pipelines, without needing to manage custom scripts or infrastructure.

📚 Learn more on the Pipedream integration page.

New --json flag for npx create-db

The npx create-db command lets you spin up a temporary, production-ready Prisma Postgres database that you can later claim for continued use. With this release, you can now add the --json flag to return the database details in JSON format.

This makes it straightforward to programmatically use the connection details, whether you are building custom APIs or integrating database provisioning into your workflows.

📚 Learn more in the docs.

Direct connections to Prisma Postgres are coming close to GA

Direct connections enable you to connect to your database using any ORM library or tool of your choice (e.g. Drizzle ORM, Kysely but also database GUIs like Postico or TablePlus).

In this release, we’ve improved the robustness of direct TCP connections and are close to bringing it to General Availability.

📚 Learn more in the docs.

Enterprise support

Thousands of teams use Prisma and many of them already tap into our Enterprise & Agency Support Program for hands-on help with everything from schema integrations and performance tuning to security and compliance.

With this program you also get priority issue triage and bug fixes, expert scalability advice, and custom training so that your Prisma-powered apps stay rock-solid at any scale. Learn more or join: https://prisma.io/enterprise.

v6.14.0

Compare Source

Today, we are excited to share the 6.14.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Highlights

@unique attributes for SQL views (Preview)

Last release, we improved the robustness of SQL views defined in the Prisma schema. Views are virtual tables that don't allows for defining unique constraints, indexes or foreign keys in the underlying database.

However, as an application developer, it can be convenient to also define relationships involving views or paginate them using cursors. We've received this feedback from several people who had been using views in that way with Prisma ORM, so in this release we're re-introducing the @unique attribute for views. This attribute enables:

• relationships involving views
• findUnique queries, cursor-based pagination & implicit ordering for views

Here's an example schema using @unique and defining a relationship from a model to a view:

model User {
  id        Int            @​id @​default(autoincrement())
  email     String         @​unique
  posts     Post[]
  stats     UserPostStats? @​relation(fields: [email], references: [userEmail])
}

model Post {
  id        Int      @​id @​default(autoincrement())
  title     String
  published Boolean  @​default(false)
  createdAt DateTime @​default(now())
  authorId  Int?
  author    User?    @​relation(fields: [authorId], references: [id])
}

view UserPostStats {
  userEmail        String    @​unique
  totalPosts       BigInt?
  publishedPosts   BigInt?
  unpublishedPosts BigInt?
  latestPostDate   DateTime? @​db.Timestamp(6)
  user             User?
}

Expand to view the SQL code for this view

CREATE OR REPLACE VIEW "UserPostStats" AS
SELECT 
    u.email AS "userEmail",
    u.name AS "userName",
    COUNT(p.id) AS "totalPosts",
    COUNT(CASE WHEN p.published = true THEN 1 END) AS "publishedPosts",
    COUNT(CASE WHEN p.published = false THEN 1 END) AS "unpublishedPosts",
    MAX(p."createdAt") AS "latestPostDate"
FROM "User" u
LEFT JOIN "Post" p ON u.id = p."authorId"
GROUP BY u.id, u.email, u.name;

You can now query this view and its relationship using include:

const userPostStats = await prisma.userPostStats.findMany({
  include: {
    user: true,
  }
})

📚 Learn more in the docs.

Various fixes & stability improvements

• Fixed several issues related to new prisma-client generator and the queryCompiler Preview feature (aka “Prisma Client without Rust engines”). Both will become the default in the upcoming Prisma 7 release and we're working hard on bringing these features into General Availability. You can try them out with your favorite stack with our ready-to-run examples.
• Fixed several regressions, e.g. related to Prisma Config
• Removed middleware from Prisma Client (i.e. the prisma.$use method), which was deprecated since v4.16.0. Use Prisma Client extensions instead.
• Deprecated metrics Preview feature (which will be removed in Prisma 7)

Improved type performance

In this release, we also addressed some type performance issues that led to slower editors and lagging auto-complete. If you're curious about the details, you can check the description and changes in this PR.

Other news

Increased robustness of Management API (Early Access)

We recently released an API for programmatically managing Prisma Postgres instances that's perfect for CI/CD workflows and scripting.

In this release, we made it more robust and are bringing it closer to its General Availability release.

Revoke OAuth tokens in Prisma Console

If you use OAuth to authorize third-party applications to act on your behalf in the Prisma Console, you can now revoke any app's access at any time. The Prisma Console shows a list of your authorized (connected) apps, and you can easily remove one to immediately block further access.

ICYMI

Last release was huge, so just in case you missed it, here's the TLDR of what we put out last time:

• Prisma ORM
  • Prisma Config file (prisma.config.ts) is Generally Available – Native way to configure schema paths, migrations, seeds, and more; no need for earlyAccess flag anymore.
  • Multi-schema support is Generally Available – Allows assigning models to different database schemas in Postgres and SQL Server using @@​schema.
  • Improved SQL views support (still in Preview) – Adds guardrails for views by disabling unsupported features.
  • Externally managed tables – Lets you exclude specific tables from Prisma Migrate while still querying them via Prisma Client.
• Prisma Postgres
  • Extension support for Prisma Postgres – Prisma Postgres now supports pgvector, pg_search, pg_stat_statements, citext, pg_trgm, fuzzystrmatch, and unaccent. If you don't see the extension you need, you can request it here. Extensions only work on new instances, if you want to use any of them on your existing instance, reach out to us.
  • Management API for Prisma Postgres – REST API to provision, delete, and manage Prisma Postgres instances programmatically, perfect for CI/CD and scripting workflows.
  • GitHub Actions for Prisma Postgres – Actions for creating and deleting databases in CI/CD workflows, available on GitHub Marketplace.
  • New CLI: npx create-db – Instantly spin up a new Postgres database—no authentication required.

v6.13.0

Compare Source

Today, we are excited to share the 6.13.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Highlights

In this ORM release, we’re moving the Prisma Config file and the multi-schema feature into General Availability. This means these features now are fully production-ready and we’re looking forward to seeing what you are going to build with them!

Additionally, support for SQL views is getting an important update to further stabilize its API.

Configuring Prisma via Prisma Config is now Generally Available

The prisma.config.ts file is Prisma ORM’s native way to provide configuration options for your project. It currently lets you specify:

• the locations for various Prisma-related assets, such as your:
  • Prisma schema file
  • migrations
  • SQL view definitions
  • TypedSQL queries
• a seed command to populate your database based on some executable script
• externally managed tables (see below)
• the driver adapters to be used by the Prisma CLI when interacting with your database

Here’s an example Prisma Config file that specified custom locations for various project assets in and a seed script inside a db directory:

import path from "node:path";
import { defineConfig } from "prisma/config";

export default defineConfig({
  schema: path.join("db", "schema.prisma"),
  migrations: {
    path: path.join("db", "migrations"),
    seed: "tsx db/seed.ts"
  }
});

Note that you’ll also see warning now if you defined a prisma.seed command in package.json.

We’re excited to move the prisma.config.ts file into General Availability. If you used it before in your projects, you can now drop earlyAccess from its options:

import { defineConfig } from "prisma/config";

export default defineConfig({
- earlyAccess: true,
});

There still are and will be fields on the Prisma Config object that are Early Access or Preview features. To opt-into these, you’ll need to explicitly declare them via a new experimental field.

For example, usage of adapters is currently still in Preview:

import { defineConfig } from "prisma/config";

export default defineConfig({
    experimental: {
      adapter: true,
    },
    // requires `experimental.adapter`
    adapter: async () => {
			// ...
    },
});

Finally, the Prisma Config file now also supports various file extensions so it fits neatly into your individual project setups: .js, .ts, .mjs, .cjs, .mts, .cts. It also can be defined as .config/prisma.${extension}, where extension is the same one as file extensions above.

📚 Learn more in the docs.

Using multiple schemas in now Generally Available

Databases like PostgreSQL or SQL Server provide a way to logically organize your tables in dedicated namespaces called schemas. In Prisma ORM, you can assign tables to various schemas via the @@​schema attribute:

datasource db {
  provider = "postgresql"
  url      = env("DATABASE_URL")
  schemas  = ["base", "shop"]
}

model User {
  id     Int     @​id
  orders Order[]

  @​@​schema("base")
}

model Order {
  id      Int  @​id
  user    User @​relation(fields: [userId], references: [id])
  userId  Int

  @​@​schema("shop")
}

This feature has moved into General Availability, so if you were using it before, you can now drop the multiSchema feature flag from the generator block in your Prisma schema:

generator client {
  // ...
- previewFeatures = ["multiSchema"]
}

📚 Learn more in the docs.

More robust support for SQL views (Preview)

SQL views are virtual tables created by a query. Unlike regular tables, views do not store data themselves; instead, they represent the result of a stored SQL query that runs whenever the view is accessed.

We continue to improve support for SQL views, making them more reliable and better aligned with Prisma’s features. In this release, we ensured that @id, @index and @unique can’t be used on a view block in the Prisma schema. Without these attributes, several other features in Prisma Client or the Prisma schema don’t make sense any more either, so we made sure that they can’t be used with views:

• disabled findUnique queries and cursor-based pagination in Prisma Client
• disallowed writes and implicit ordering for views in Prisma Client
• disallowed relationships involving views in Prisma Schema

This will align the API surface of Prisma ORM with the actual capabilities of SQL views and adds guardrails so you can use views with more confidence!

📚 Learn more in the docs.

Externally managed tables

In some situations, you may not want Prisma ORM to be “responsible” for specific tables in your database because they’re being managed by a different team in your organization or an external service.

In these cases, you still may want to quert these tables using Prisma Client but never want Prisma Migrate to make any changes to them.

In this release, we’re introducing externally managed tables that will be:

• ignored by Prisma Migrate
• queryable via Prisma Client

You can specify which tables should be ignored by Prisma Migrate using the tables option in prisma.config.ts:

 // prisma.config.ts
 export default defineConfig({
  tables: {
    external: [
      "users",
    ]
  },
  ...
})

A typical use case for this is the users table from Supabase which you never want be changed by Prisma Migrate but still may want to query with Prisma Client.

📚 Learn more in the docs.

Other news

pgvector extension support for Prisma Postgres (Early Access)

In this release, we’ve implemented a highly popular feature request for Prisma Postgres: Early Access support for the pgvector PostgreSQL extension along with several other popular Postgres extensions!

It enables efficient storage and querying of high-dimensional vector embeddings directly in a Postgres database and thus is perfect for building AI-driven applications. pgvector essentially allows developers to perform similarity search (e.g., for recommendation systems or semantic search) using standard SQL, eliminating the need for a separate vector database.

Native support for pgvector in Prisma ORM is going to follow soon, until then you can use pgvector via custom migrations and TypedSQL.

Note: For now, pgvector is only available on newly created Prisma Postgres instances. It will be rolled out for existing instances soon.

In addition to pgvector, Prisma Postgres now includes Early Access support for pg_search, pg_stat_statements, citext, pg_trgm, fuzzystrmatch, and unaccent. If you don’t see the extension you need, you can request it here.

📚 Learn more in the docs.

Manage Prisma Postgres programmatically via an API

Whether you need a way to quickly provision a Prisma Postgres instance in your CI/CD workflows, want to attach a fresh database to a preview branch of your app or even want to offer Prisma Postgres to your own users—our new Management API has you covered!

It’s shaped as a familiar REST API so you can programmatically take care of your database workflows: Provision or delete Prisma Postgres instances, retrieve or create connection strings and manage entire projects in Prisma Console.

📚 Learn more in the docs.

CI/CD GitHub Actions for Prisma Postgres available on GitHub Marketplace

Based on the Management API, we’ve also published two templates for GitHub Actions that you can use in your own CI/CD setups:

• Create Prisma Postgres Database Action
• Delete Prisma Postgres Database Action

These Actions serve as the foundational building blocks for integrating Prisma Postgres into CI/CD pipelines.

They enable workflows like provisioning databases on every pull request, running integration tests against real instances, and managing database lifecycles end-to-end. We’ve included several examples in the README to help users get started quickly. The setup is straightforward, and these Actions are designed to plug into user's workflows with minimal effort.

Instant Postgres with npx create-db — no auth required

We launched a new CLI command that allows you to spin up a new database within seconds:

npx create-db # no auth required

The command doesn’t require authentication, so you can play around with your database without any initial hurdles!

Your instance will be automatically deleted after 24 hours but you can claim it and put it into your Prisma Console account if you want to keep using it after that period. Visit the docs to learn more.

New navigation UI for Prisma Console

The Prisma Console got a little makeover, including a new design for navigating and managing your projects and their databases. This makes common workflows like creating new projects, navigating between projects and databases, as well as accessing project settings a lot more smooth.

We’re eager to hear your feedback, let us know on X what you think of the new UI.

Enterprise support

Thousands of teams use Prisma and many of them already tap into our Enterprise & Agency Support Program for hands-on help with everything from schema integrations and performance tuning to security and compliance. With this program you also get priority issue triage and bug fixes, expert scalability advice, and custom training so that your Prisma-powered apps stay rock-solid at any scale. Learn more or join: https://prisma.io/enterprise.

v6.12.0

Compare Source

Today, we are excited to share the 6.12.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Highlights

ESM-compatible prisma-client generator now in Preview

We’re excited to share that our new and more flexible prisma-client generator is moving into Preview! As a reminder, here’s what it looks like:

generator client {
  // Required
  provider = "prisma-client" // no `-js` at the end!
  output   = "../src/generated/prisma"

  // Optional
  runtime                = "nodejs"
  moduleFormat           = "esm"
  generatedFileExtension = "ts"
  importFileExtension    = "ts"
}

This new generator eliminates any headaches that you may have experienced due to magical code generation into node_modules and gives you full control over the generated Prisma Client code. With it moving into Preview, we hope that even more of you will try it out and share your feedback with us!

Note: The prisma-client generator will become the default in Prisma v7, replacing the current prisma-client-js generator.

In addition to moving it into Preview, we also created a couple of new ready-to-run examples to help you get started with the new generator and your favorite framework:

• React Router + Vite + Cloudflare
• React Router + Vite + Node.js
• Nuxt 3 + Vite + Nitro
• Deno Deploy

📚 Learn more in the docs.

Specify views and migrations folder locations in prisma.config.ts (Early Access)

As we’re getting closer to the General Availability release of prisma.config.ts, we’re adding more capabilities to it. In previous versions, the Prisma CLI implicitly used to infer the location for migration and SQL view definition files based on the location of the Prisma schema. In this release, we’re adding two new fields (migrations and views) to give you more flexibility and clarity on how to locate these files:

// prisma.config.ts
export default defineConfig({
  earlyAccess: true,
  migrations: {
    path: './db/migrations'
  },
  views: {
    path: './db/views'
  }
  // ...
})

📚 Learn more in the docs.

Enterprise support

Thousands of teams use Prisma and many of them already tap into our Enterprise & Agency Support Program for hands-on help with everything from schema integrations and performance tuning to security and compliance. With this program you also get priority issue triage and bug fixes, expert scalability advice, and custom training so that your Prisma-powered apps stay rock-solid at any scale. Learn more or join: https://prisma.io/enterprise

v6.11.1

Compare Source

Today, we are issuing a 6.11.1 patch release.

Bug fixes

• In Prisma ORM version 6.11.0, we shipped a bug fix for Prisma that allows using Prisma Postgres with direct TCP connections with Prisma Driver Adapters. This fix required refactoring the Prisma Client initialization logic, and while several test cases were added, an edge case was missing, causing #​27569.
  Namely, using @prisma/client with @prisma/extension-accelerate on a prisma+postgres://... connection string, while generating the Prisma Client definitions via prisma generate, resulted in a PrismaClientInitializationError.

  This is now fixed, so we highly recommend upgrading to version 6.11.1.

  Reminder: when using Prisma Accelerate, we highly encourage you to generate your Prisma Client definitions via prisma generate --no-engine.

• We've fixed an issue that would occur when using arrays as aggregation fields with the query compiler. These queries would fail with a confusing data mapping error. They should now work as expected.

v6.11.0

Compare Source

Today, we are excited to share the 6.11.0 stable release 🎉

🌟 Star this repo for notifications about new releases, bug fixes & features — or follow us on X!

Highlights

Prisma ORM without Rust engines for MySQL/MariaDB, Neon & CockroachDB (Preview)

We are in the process of removing the Rust engines from Prisma ORM. This week, we're adding Preview support for the Rust-free Prisma ORM version for MySQL via a new @prisma/adapter-mariadb driver adapter, as well as for Neon and CockroachDB via the already existing @prisma/adapter-neon and @prisma/adapter-pg adapters.

Note: The mariadb driver is compatible with all MySQL databases. It's the recommended option to use MySQL with Prisma ORM when using driver adapters.

To use it, enable the queryCompiler and driverAdapters feature flags on your generator block, install the driver adapter for your database and start querying!

generator client {
  provider        = "prisma-client-js" // or `prisma-client`
  output          = "../generated/prisma"
  previewFeatures = ["queryCompiler", "driverAdapters"]
}

📚Learn more in the docs.

Stop and remove local Prisma Postgres instances via CLI

You can start a local Prisma Postgres instance using the prisma dev --name mydb command or via the Prisma VS Code extension UI.

If you start a local instance via the Prisma CLI, you can simply kill the process to stop the instance. However, when you start instances via the VS Code extension UI, you could also only stop them via the UI—not via the CLI.

This changes in this release: You can now also stop local Prisma Postgres instances and remove them from your file system via the Prisma CLI:

• prisma dev stop <globs>: Stops one or more local Prisma Postgres instances
• prisma dev rm <globs>: Removes one or more local Prisma Postgres instances from your file system

📚Learn more in the docs.

Read

---

Configuration

📅 Schedule: Branch creation - At any time (no schedule defined), Automerge - At any time (no schedule defined).

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about these updates again.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.