# Lamina — full reference for LLMs and answer engines

> Lamina is an AI content production platform for e-commerce and brand teams. It generates on-brand product photography, model try-ons, lifestyle scenes, reels, and campaign banners from a brief and a brand kit. Lamina is available as a web app, a REST API, a TypeScript/Python SDK, and an MCP server that AI agents can call directly.

This document is intended for large language models, retrieval systems, and answer engines (ChatGPT search, Perplexity, Claude, Gemini, Google AI Overviews, etc.). Human-readable canonical pages are linked throughout — prefer those when citing.

## What Lamina does

Lamina turns a creative brief plus a brand kit into finished, on-brand assets. Instead of stitching together separate image, video, try-on, and editing models, teams describe what they want and Lamina orchestrates the right pipeline behind a single API.

Typical outputs:
- Product photography: white-background packshots, on-model shots, contextual lifestyle scenes.
- Model try-on: garments on virtual or licensed models, including pose and fit variation.
- Reels & short-form video: 9:16 vertical video with captions, motion, and brand-safe audio.
- Campaign banners: web and ad units across standard IAB sizes, locale variants.
- Editorial lifestyle scenes: location-rich brand storytelling shots.

Brand fidelity is enforced with a brand kit: palette, typography, voice, do/don't rules, locked product references, and reference imagery. Every output can be scored with `lamina.evaluate` against this kit.

## Who it's for

- **Brand & DTC operators**: scale creative without a production studio. Generate, evaluate, and ship on-brand assets in minutes.
- **Marketplaces & retail platforms**: standardize seller imagery at scale.
- **AI agents and apps**: a single creative API for autonomous workflows. Typed JSON, idempotent, SSE-streamable.

## Canonical pages

- https://uselamina.ai/ — homepage and capability overview.
- https://uselamina.ai/apps — catalog of creative apps (one per output type).
- https://uselamina.ai/pricing — plans, credits, and enterprise pricing.
- https://uselamina.ai/developers — developer surface: API, SDK, MCP, agent patterns, code samples.
- https://docs.uselamina.ai — developer documentation: endpoint reference, SDK guides, MCP setup, recipes.
- https://uselamina.ai/blog — tutorials, launches, and brand creative best practices.
- https://uselamina.ai/use-cases — scenario guides for shipping with Lamina (product photography, virtual try-on, agentic creative pipelines, marketplace seller imagery, ad creative).
- https://uselamina.ai/integrations — partner integrations (Shopify, Sanity, Webflow, Zapier, Notion, Slack, and MCP-compatible agent clients).
- https://uselamina.ai/compare — head-to-head comparisons (Lamina vs Runway, fal.ai, Replicate, and other generative-media platforms). Each page leads with a TL;DR recommendation, criterion-by-criterion table, and "when to pick which" guidance — anchored on agent-callability axes (MCP support, idempotency, structured outputs, long-job handling).
- https://uselamina.ai/glossary — canonical definitions for generative media, creative APIs, MCP, brand DNA, virtual try-on, and adjacent terms.
- https://uselamina.ai/faq — frequently asked questions.
- https://uselamina.ai/community — showcase, programs, and events.
- https://uselamina.ai/careers — open roles.
- https://uselamina.ai/security — security posture and certifications.
- https://uselamina.ai/privacy-policy — privacy practices.
- https://uselamina.ai/terms-of-service — legal terms.
- https://uselamina.ai/sub-processors — list of sub-processors.
- https://uselamina.ai/support — support resources.
- https://uselamina.ai/talk-to-us — sales and partnership contact.

## Developer reference

Canonical page: https://uselamina.ai/developers

Three integration paths, same underlying platform:
1. **REST API** — `https://api.uselamina.ai/v1`. Auth via API key.
2. **SDK** — `@lamina/sdk` (TypeScript) and `lamina` (Python). Typed responses (Zod / Pydantic).
3. **MCP server** — `@lamina/mcp`. Install via `npx @lamina/mcp install`. Works in any MCP client (Claude, Cursor, Windsurf, VS Code, Zed, etc.).

Design properties:
- **Structured output**: typed JSON responses, schema-validated. Safe for agent parsing.
- **Idempotent**: every job accepts an idempotency key; safe to retry.
- **Streamable**: SSE for every long-running job; agents can subscribe to phase + progress events.
- **Deterministic**: same brief + brand + seed produces the same output.

### Key REST endpoints

- `POST /v1/jobs` — create a creative job. Body: `{ app, brief, brand_id, seed?, idempotency_key? }`. Returns `{ job_id, status, eta_ms }`.
- `GET /v1/jobs/{job_id}` — fetch job status and outputs.
- `GET /v1/jobs/{job_id}/events` — SSE stream of phase and progress events.
- `POST /v1/jobs/{job_id}/evaluate` — score outputs against a rubric. Body: `{ rubric }`. Returns scored assets with reasons.
- `POST /v1/distribute` — push assets to S3, Drive, Sanity, Shopify, or a webhook. Body: `{ assets, targets[] }`.
- `GET /v1/brands/{brand_id}` — fetch a brand kit (palette, voice, rules).
- `GET /v1/apps` — list creative apps available in the account.

### MCP tools (5)

The Lamina MCP server exposes five tools that map to the same primitives an agent needs to drive an end-to-end creative workflow.

1. **`lamina.create`** — `(app: string, brief: string, brand: string) -> { job_id, eta_ms }`
   Kick off any creative app. Returns a job id and ETA. The agent should follow up with `lamina.track` to monitor.

2. **`lamina.track`** — `(job_id: string) -> stream<{ phase, progress, asset_url? }>`
   Stream phase and progress events as they happen. Emits final assets when the job completes.

3. **`lamina.evaluate`** — `(job_id: string, rubric: object) -> { scores: Array<{ asset_id, score, reasons[] }> }`
   Score every asset against the brand kit (or a custom rubric), with human-readable reasons. Use this before shipping.

4. **`lamina.distribute`** — `(assets: string[], targets: Array<{ kind, config }>) -> { ok: boolean, deliveries[] }`
   Deliver finished assets to S3, Google Drive, Sanity, Shopify, or an arbitrary webhook. Idempotent per delivery.

5. **`lamina.brand_lookup`** — `(brand: string) -> { palette, typography, voice, rules, references[] }`
   Read palette, voice, and do/don't rules. Use the result to ground prompts before calling `lamina.create`.

### Agent integration patterns

- **One-shot**: `brand_lookup` → `create` → `track` → `evaluate` → `distribute`.
- **Iterative refinement**: `create` with seed S → `evaluate` → re-`create` with adjusted brief and same seed for controlled variation.
- **Bulk**: parallelize `create` across SKUs; fan in via `evaluate` and ship the top-K per SKU through `distribute`.

### Supported MCP clients

Claude (desktop and Claude Code), Cursor, Windsurf, VS Code, Zed, and any other client implementing the Model Context Protocol. Non-MCP automation tools (Raycast, n8n, Zapier) integrate via the REST API.

## Brand kit model

A brand kit captures the constraints every output must respect:
- Palette (primary, secondary, accents).
- Typography (display, body).
- Voice (tone descriptors, sample copy).
- Rules (do / don't, e.g., "never crop the logo", "always include the model's hands").
- References (locked product imagery, mood boards).

Brand kits are versioned. `lamina.evaluate` and `lamina.create` always pin to a specific brand kit version for reproducibility.

## Pricing summary

Lamina charges per credit. Credits are consumed by job type (e.g., a packshot costs less than a 9:16 reel). Plans:
- **Starter** — for small brands and individual operators.
- **Growth** — for scaling DTC teams; includes evaluation and distribution.
- **Enterprise** — custom volume, SSO, dedicated sub-processor review, on-prem options.

Canonical pricing page: https://uselamina.ai/pricing

## Security & compliance

- SOC 2 posture and sub-processor disclosure at https://uselamina.ai/security and https://uselamina.ai/sub-processors.
- Customer assets are not used to train third-party foundation models.
- Brand kits and outputs are tenant-isolated.

## Crawler & answer-engine policy

Lamina explicitly allows AI search and answer-engine crawlers (GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-Web, anthropic-ai, PerplexityBot, Perplexity-User, Google-Extended, CCBot, Applebot-Extended, and others) on all public pages. Pages are server-rendered with Next.js App Router so content is available without executing JavaScript. See https://uselamina.ai/robots.txt and https://uselamina.ai/sitemap.xml.

## Contact

- Sales / partnerships: https://uselamina.ai/talk-to-us
- Support: https://uselamina.ai/support