Prisma (@prisma)

Pinned Post

Prisma

@prisma

Jun 10

Agents changed how fast we can build software But shipping still feels slower than it should, because your app, database, deploys, logs, and previews all live in different places Prisma Compute is our next step toward bringing that whole loop together

0:33

26,053

Prisma reposted

Tyler Hogarth @tylerhogarth

I've opted for using the Prisma Stack for everything I build: 1. Prisma (typescript, prisma next, postgres, bun). 2. RR7 (or whatever you prefer) My deployed setup is as now easy as serving localhost... prisma.io/stack

Prisma Stack: a complete TypeScript foundation | Prisma

Prisma Compute, Prisma Postgres, Bun and TypeScript. A complete, opinionated stack with room for your frontend.

prisma.io

697

Prisma reposted

Kristóf Siket @kristofsik

Jul 2

New @prisma Compute command of the day: prisma-cli build logs <build_id> It works with a --follow flag if you want a live tail. Additionally, if you are using Compute through GitHub, your agent can fetch & run it from the Check output.

0:18

1,886

Prisma reposted

Kristóf Siket @kristofsik

Jul 1

We shipped a --no-promote flag for prisma app deploy to add more control over deploy promotion. Fun fact: we're its first users, for one of our services that takes some time to boot. 🧪 The flow: - prisma app deploy --no-promote - health-check the candidate - prisma app promote

1,111

Prisma

@prisma

Jun 26

highlights from Prisma in the last week. more in the 🧵

2,986

Prisma

@prisma

Jun 26

Prisma Next will soon support customizing many-to-many relations by defining an explicit junction model with arbitrary fields

763

Prisma

@prisma

Jun 26

Prisma next will also have more intuitive matching of forward and backward relations in cases where they can’t be matched automatically due to ambiguity (explicit inverse fields instead of relation names)

658

Prisma

@prisma

Jun 25

👉 pris.ly/compute-pricing-tl

Price the Work, Not the Workflow

Agents can ship, preview, test, and retry in loops. Infrastructure pricing should charge for the work an app performs, not the workflow around it.

prisma.io

Luan

@luan_vdw

Jun 25

This is how we think infrastructure pricing should evolve as agents take a bigger role in how software is built. With agents, development runs in tighter, faster loops. More previews, retries, tests, branches, deploys. Pricing has to separate the factory activity from the actual application work. That should make builders feel more capable, not more exposed.

2,557

Prisma

@prisma

Jun 24

🇩🇪 Berlin friends, come hang out with us in our office tomorrow at the @typescript Berlin meetup. We’ll have talks on TypeScript architecture, plus pizza, drinks, and good company.

2,592

Prisma

@prisma

Jun 24

Join our meetup group here 👇 pris.ly/ts-meetup

TypeScript Berlin | Meetup

A community for developers interested in TypeScript and JavaScript frameworks. We get together to share knowledge, explore real-world problem solving, and discuss the things we're building.

meetup.com

1,192

Prisma

@prisma

Jun 24

🫶 @kapa_ai has been great at helping users get answers from our docs. now it also helps the Prisma MCP answer questions directly 👇 pris.ly/kapa-mcp

Search the Prisma Docs Using Your Coding Agent

The Prisma MCP server now answers documentation questions in your editor. Ask about ORM, Postgres, or Compute and get a cited answer, no tab-switching, no extra setup.

prisma.io

Ankur Datta

@ankur_datta_007

Jun 24

most Prisma questions i’ve answered over the past couple of years can be solved with the ask ai feature in our docs (built with @kapa_ai). and if you’re using the prisma mcp, you can ask the same questions directly in your agent and get answers right from the terminal.

2,471

Prisma reposted

Kristóf Siket @kristofsik

Jun 23

"Loops" is clearly the new buzzword. Easy when the flow is local, but when your deployment happens via GitHub, how does your agent know it failed? @prisma Compute attaches an automated check to every commit. Your agent can fetch it via `gh` CLI and get the full build log.

0:12

3,160

Prisma

@prisma

Jun 23

Build with Prisma Compute and tag us. also, checkout what @rtbenfield built 👇 pris.ly/gridlock-tyler

0:04

Tyler Benfield

@rtbenfield

Jun 23

Wait you guys didn't have websockets? How else would you build something this cool? Real-time light cycle race with a simple in-memory session and player queue running on Prisma Compute. Have fun✌🏻 gridlock.rtbenfield.dev/

2,809

Prisma

@prisma

Jun 22

look at the work our team is doing to help engineers ship faster without lowering the bar 👇

Tyler Hogarth @tylerhogarth

Jun 22

A skill should define a contract, not narrate a process. What goes in. What comes out. What standard it has to meet. A skill shouldn't read like a blog post. It should read like an interface. prisma.io/blog/evolving-agen…

3,482

Prisma

@prisma

Jun 19

blog 👉 pris.ly/bun-image

Søren Bramer Schmidt

@sorenbs

Jun 19

Because @prisma Compute runs on Bun, you can take advantage of Buns powerful functionality for tasks you'd otherwise have to configure an external service for. Check out how to use Bun.Image to generate thumbnails, and more: prisma.io/blog/image-transfo…

7,563

Prisma reposted

Kristóf Siket @kristofsik

Jun 19

At @prisma, we are naturally quite heavy users of Prisma Compute. Yesterday, I needed to roll back a deployment and didn't find a Console action to do so. Today, we shipped the "Promote to live" button on each eligible version on Console. Stay safe & keep on deploying!

0:10

2,671

Prisma reposted

Prisma

@prisma

Jun 17

🚀 Prisma Next performs close to raw `pg`. It reaches ~90% of raw `pg` peak throughput in our benchmark, while shipping a 148.5 KB gzipped bundle.

0:10

137

20,693

Prisma

@prisma

Jun 17

🚀 Prisma Next performs close to raw `pg`. It reaches ~90% of raw `pg` peak throughput in our benchmark, while shipping a 148.5 KB gzipped bundle.

0:10

137

20,693

Prisma

@prisma

Jun 17

Try Prisma Next today by running "bunx create-prisma@next" and read the full performance deep dive 👇 pris.ly/pn-performance

Prisma Next Is ~90% As Fast as Raw PG

Prisma Next performance benchmarks achieves ~90% of the raw pg driver's peak throughput, holds latency low under load, and ships as a 148.5 KB gzipped bundle for serverless and edge workloads.

prisma.io

1,349