ProductApr 27, 2026

Introducing the Lamina Developer Platform

Today we're opening up the platform that powers Lamina to every developer and every agent. One API, dozens of apps, no model wrangling.

Jophin Joseph

Jophin Joseph

Agent Engineering

Lamina Developer Platform launch announcement

Today we're launching the Lamina Developer Platform — the same surface that powers Lamina, now open to every developer and every AI agent. One API, dozens of generative apps, native MCP tools, and outputs you can ship into production.

Why we built it

Generative media is exploding, but shipping it is still painful. Teams stitch together five model providers, three image pipelines, a captioning service, an evaluator, and a publishing layer — then spend the next quarter babysitting prompt regressions and aspect-ratio bugs. The work that should be describe what you want, get usable output becomes weeks of plumbing.

We built the Developer Platform around a different premise: you should call apps, not models. Each app is an opinionated pipeline — try-on, brand-consistent product shots, lifestyle video, lip-synced UGC — that already knows which model to use, how to prompt it, how to evaluate the result, and how to format it for the channel you care about. You bring intent. We bring the stack.

What's in the platform

From day one, the platform exposes the full Lamina pipeline through a single, agent-native API: CreateTrackEvaluateDistribute. Every step is addressable, every artifact is inspectable, and every run is reproducible.

Call apps, not models

Here's a complete brand-consistent product shot — kept on-brand, evaluated, and ready for distribution — in a single call:

create.tstypescript
import { Lamina } from '@lamina/sdk'

const lamina = new Lamina({ apiKey: process.env.LAMINA_API_KEY })

const run = await lamina.apps.run('product-shot', {
  brandKit: 'acme-ss26',
  product: { sku: 'SNK-1043', image: 'https://cdn.acme.com/sku/1043.png' },
  scene: 'studio-soft-light',
  aspect: '4:5',
})

await run.waitForReady()
console.log(run.assets[0].url)
TypeScript SDK — same surface available in Python and via raw HTTP.

Built for agents from day one

Most generative APIs treat agents as an afterthought. We did the opposite. Every app is exposed as an MCP tool with structured inputs, machine-readable outputs, and self-describing schemas — so an agent can plan a campaign, run apps, evaluate results, and iterate without you writing a single line of glue code. If you've used Claude, Cursor, Windsurf or VS Code with MCP, you're already five minutes from generating production assets.

Get started today

Grab an API key, install the MCP server, or paste a curl into your terminal — whichever path is shortest for you. We're rolling out new apps every week, and the pricing is built around what you actually generate, not seats.

We can't wait to see what you build. Tell us in #dev on the community Slack — or, better, ship something and let your agent post about it.

Tagslaunchdeveloper-platformapimcpagents