Store files with Object Store buckets and see your workspace at a glance

Prisma projects can now store files. Object Store buckets are S3-compatible storage that lives next to your Prisma Postgres databases and Prisma Compute apps, managed from the Prisma Console, the Management API, or MCP tools.

The Prisma Console workspace page is now an overview dashboard: usage against your plan, the current invoice, deploy health, and your most active projects on one screen.

Prisma Next 0.16.0 is on npm, and bringing an existing database to Prisma Next is now much smoother. The prisma.io docs are also rebuilt for coding agents, with per-section llms.txt indexes and a markdown version of every page.

Highlights

NewStore files in your project with Object Store buckets

Apps that store user uploads or generated files needed a separate storage vendor, with its own account, credentials, and billing. Every Prisma project can now create Object Store buckets: S3-compatible storage that lives alongside your databases and apps. Create a bucket in the Prisma Console, generate an access key (the secret is shown exactly once, so store it safely), and read and write objects with any S3 client or SDK.

Buckets can also be managed through the new /v1/buckets Management API endpoints and matching MCP tools, so agents can provision storage the same way they provision databases. Deleting a bucket removes its contents in one step, even when it is not empty.

NewSee your whole workspace at a glance

The Console landing page used to show a usage dashboard, so answering "is everything healthy?" meant clicking through projects. The workspace page is now an overview: billing-cycle usage with your included-limit line, the current invoice with per-line meters, and your top projects with their deploy health, repository, latest commit, and share of usage. The detailed usage dashboard, including usage-threshold and database-quota alerts, moved to its own Usage page, and the projects list is now paginated.

NewPrisma Next 0.16.0

Pointing Prisma Next at an existing database could produce a schema its own toolchain rejected: contract infer printed output that contract emit refused, index types like gin failed to emit, and a column's exact type and its auto-update behavior could not be declared together. In 0.16.0, inferred schemas work as generated: infer output round-trips cleanly through emit, the built-in Postgres index types (gin, gist, brin, and friends) emit, and new temporal presets let one field declaration carry both the column type and its auto-update behavior:

model Page {
  updatedAt temporal.timestamp(3, onCreate: now, onUpdate: now)
  lastSeen  temporal.timestamp(3)
  touched   temporal.timestamptz(onUpdate: now)
}

Note: Prisma Next is in Early Access. Scope and behavior may still change. This release includes breaking changes, listed below.

Read the v0.16.0 release notes for the full list. Shipped in prisma/prisma-next#1003, prisma/prisma-next#1011, and prisma/prisma-next#1019.

Prisma Next

Beyond the 0.16.0 headline changes, your editor now gives live schema feedback, and polymorphic queries get more capable.

New

Schema problems show up in your editor as you type. The language server now reports unresolvable relations, type and codec mismatches, and extension-block errors live, instead of waiting for contract emit. If your config file fails to load, the error appears on the config file itself, and existing diagnostics stay put instead of disappearing. (prisma/prisma-next#972, prisma/prisma-next#974)

Improved

Polymorphic queries can follow relations declared on a variant: narrow a query with .variant(...) and .include() the relations that variant declares. (prisma/prisma-next#976)

Fixes and improvements (4)

Fixed

A Bytes column selected inside .include() now returns bytes instead of raw \x... hex text. (prisma/prisma-next#942)
Migrations run their operations in dependency order, fixing cases such as a column being dropped before its own constraint. (prisma/prisma-next#992)
Count-returning updates and deletes on a narrowed variant no longer generate SQL that references a missing table. (prisma/prisma-next#940)
More contract infer fixes: Decimal fields connect without a codec error, dbgenerated(...) defaults no longer report permanent drift in db verify, and foreign keys that point outside the introspected tables are explained in the output instead of silently dropped. (prisma/prisma-next#1011)

Prisma Compute

App detection moved into the public SDK, and app URLs are now correct before your first deploy.

New

The Prisma Compute SDK can inspect a repository without cloning it: detectComputeApp reads a repository snapshot and reports the app's framework, build type, HTTP port, and entrypoint. Elysia, Express, and Fastify apps are recognized and run on Bun. Available since @prisma/compute-sdk 0.37.0.

Improved

The GitHub repository picker in the Prisma Console loads repositories page by page, so installations with more than 100 repositories can reach all of them, and switching accounts mid-load no longer briefly shows the wrong account's list.

Fixed

The app URL returned before your first deployment now names the domain your app will actually serve on. Previously it could point at the wrong region and never resolve.

Prisma Postgres

One reliability improvement for databases waking from idle.

Improved

Cold starts no longer run through a single per-region coordinator, so one degraded coordinator cannot block databases from waking across its region.

Prisma docs

The prisma.io docs are now built for coding agents.

New

The root llms.txt is a compact index linking ten per-section indexes instead of one file too large for agents to read, every page carries a crawler-visible pointer to its markdown version, and MCP server discovery is published. Point your agent at prisma.io/docs and every page is reachable from the root index. (prisma/web#8087)
A Plasmic integration guide walks through connecting a Plasmic-built React app to Prisma, with queries configured visually in Plasmic Studio and credentials staying in your app's environment. (prisma/web#8095)

Breaking changes

⚠️ Breaking: Prisma Next 0.16.0 changes the emitted contract shape. Foreign keys and their backing indexes are now separate contract entries, a 1
relation over a non-unique foreign key fails emit with PSL_NON_UNIQUE_BACKRELATION, contract infer declares identity-column defaults, and inferred back-relation names no longer double-pluralize. Your database schema is unaffected. To migrate, re-run contract emit and follow the 0.15-to-0.16 upgrade recipe. (prisma/prisma-next#989, prisma/prisma-next#1015, prisma/prisma-next#1011)

⚠️ Breaking: The deprecated experimental /v1/compute-services and /v1/versions Management API endpoints are removed. To migrate, use the /v1/apps and /v1/deployments endpoints, which cover the same operations.

Guides and articles

Extending Prisma Next with Typed Postgres ltree: use PostgreSQL ltree in Prisma Next with prisma-ltree, with typed path columns plus ancestor and descendant queries.
You Don't Need Elasticsearch, Postgres Already Has Full-Text Search: build full-text search with tsvector, GIN indexes, and pg_trgm typo tolerance on Prisma Postgres.
How to Make Your Docs Agent-Ready: coding agents fail silently on truncated indexes, broken links, and missing metadata. What we fixed in the Prisma docs, and how to audit yours.

Need help applying these changes in production? Prisma Enterprise Support can help with schema design, performance, security, and compliance.