37 MCP tools · REST API · Free in beta

The complete AI video studio,
in your AI editor.

Playcut ships a Model Context Protocol (MCP) server plus a REST API. Every operation in the Playcut studio — build a custom AI actor, render an actor shoot, generate a Veo cinematic clip, clone a voice, manage a brand kit, batch 50 variants — is exposed as a tool any AI agent can call. Wire it into Claude, ChatGPT, or Cursor in 60 seconds.

Get your API key 60-second setup

Free during open beta · No usage tier · Pay only for credits per generation

What MCP unlocks

Generate from anywhere. Ship from your editor.

Ad volume, on tap

Spawn 50 UGC ad variants in a single Claude prompt. Same actor, 5 languages, 10 hooks. Your performance team stops being the bottleneck.

↳ 50 variants · ~14 minutes · ~$30 in credits

Brand consistency, enforced

Pass a brand-kit ID into every call. Palette, typography, voice, logo placement, and your custom actor's face all stay locked across every variant.

↳ 100% character consistency · brand-kit-aware

Studio cost, gone

Skip the studio booking, the model release fees, the producer, the post house. A typical product catalog of 100 SKUs renders for under $200 in credits.

↳ ~90% cheaper than studio shoots

60-second setup

Three installs. Pick yours.

Generate an API key in your Playcut workspace settings, then drop one config block into your client. That's it.

Claude Desktop

macOS + Windows · 1-click .mcpb install or manual config

~/Library/.../claude_desktop_config.json

// claude_desktop_config.json
{
  "mcpServers": {
    "playcut": {
      "command": "npx",
      "args": ["-y", "@playcut/mcp"],
      "env": { "PLAYCUT_API_KEY": "pk_live_…" }
    }
  }
}

Claude Code

CLI agent runtime · per-project .mcp.json

<project>/.mcp.json

// .mcp.json
{
  "mcpServers": {
    "playcut": {
      "command": "npx",
      "args": ["-y", "@playcut/mcp"],
      "env": { "PLAYCUT_API_KEY": "pk_live_…" }
    }
  }
}

Cursor · Windsurf · Zed

Identical MCP config; the spec is universal

Settings → MCP

// mcp.json (Cursor) — same shape for Windsurf + Zed
{
  "mcpServers": {
    "playcut": {
      "command": "npx",
      "args": ["-y", "@playcut/mcp"],
      "env": { "PLAYCUT_API_KEY": "pk_live_…" }
    }
  }
}

ChatGPT · n8n · Zapier · custom backend

Not MCP-aware? Use the REST API directly.

api.playcut.ai

# Example: generate an image from text
curl -X POST https://api.playcut.ai/v1/generate-image-from-text \
  -H "Authorization: Bearer pk_live_…" \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Editorial portrait of …", "aspectRatio": "4:5"}'

Complete tool catalog

37 tools. Eight categories.

Every workflow surface the Playcut app exposes is reachable from MCP and REST. Build, list, render, extend, clone, share, delete — your agent has the same powers a logged-in operator does.

AI Actors

8 tools

Build, list, update, and direct custom AI actors. The Playcut Actor Engine holds 100% character consistency across stills, motion video, UGC ads, and on-product compositing — your agent can choreograph the same face across hundreds of shots in a single prompt.

generate-actor

Create a new custom AI actor from a structured description (sex, age, ethnicity, hair, body type, signature accessories). Returns an actor ID reusable across every subsequent shoot.

list-actors

List the workspace's actors plus optional curated template actors. Use the returned IDs in actor-shoot, actor-video, and actor-act calls.

actor-shoot

Render still images of a specific actor in a scene — product holds, lookbook stills, brand-kit-aware compositions. The face stays consistent across every shoot.

actor-video

Generate motion video featuring your actor — lookbook reels, demo videos, talking-head UGC. Same face, same body, ad-native vertical or horizontal.

actor-act

Direct an actor's performance — script, emotion, scene action — for higher-fidelity narrative video.

actor-video-extension

Extend an existing actor video clip — append additional seconds while preserving identity and motion continuity.

clone-actor-template

Clone an existing actor as a reusable template, useful for spinning up variants without losing the canonical face.

update-actor

Modify an existing actor's attributes — hair, outfit, recognition anchors — without breaking past renders.

Images

2 tools

Text-to-image and reference-style image generation across Imagen, Gemini, and curated fal.ai models. Your agent picks the right backend automatically based on the prompt.

generate-image-from-text

Generate an image from a text prompt using Google Imagen / Gemini. Supports 1:1, 4:5, 3:2, 16:9, 9:16, 21:9, plus grid formats (2×2, 3×3, 4×4) for batch comparison.

generate-image-from-reference

Generate an image conditioned on one or more reference images — style transfer, product compositing, or matching an existing brand-asset.

Videos

5 tools

Veo cinematic motion plus reference-style, interpolation, and extension. Your agent can ship a 30-second branded sequence from a single CSV row.

generate-video-from-text

Text-to-video with Google Veo. Anamorphic, dolly, tracking, drone aerials, long-shot coherence. 1080p+ at multiple aspect ratios.

generate-video-from-image

Animate a still you already love — image-to-video at any supported aspect ratio. Useful for turning a brand photo into a launch reel.

generate-video-from-reference

Reference-style video — lock visual language across an entire campaign by anchoring on a single reference frame.

generate-video-extension

Extend any clip seamlessly. Append motion to a finished cut without re-rendering the original frames.

generate-video-interpolation

Interpolate between two keyframes — generate the connective motion for a stop-motion or storyboard pipeline.

Voice & Speech

4 tools

The Playcut Voice Engine covers TTS in 30+ languages, voice cloning from a short audio sample, and custom voice design. Pair it with actor-video for end-to-end localized UGC ads.

generate-tts

Text-to-speech in 30+ languages with native intonation. Pair with actor-video for perfect lip-sync in localized variants.

clone-voice-from-audio

Clone a voice from a short audio sample. The cloned voice becomes reusable across every TTS call in the workspace.

design-voice

Design a brand-new synthetic voice from a textual description (pitch, energy, accent, persona). Save it to your library.

list-voices

List the workspace's voice library — cloned voices, designed voices, and the curated default set.

Assets & Files

6 tools

Every generation lands in your workspace as an asset. Your agent can list, fetch, share, move, and download — the same operations the app surface offers, exposed as MCP calls.

get-asset

Fetch full metadata + signed URL for any asset by ID. Returns dimensions, mime-type, generation lineage.

list-assets

Paginated asset listing with filters by type, folder, status, and date.

download-asset

Convenience download — given an asset ID, write the bytes to a local file path. Useful for feeding outputs into ffmpeg, editors, or other agents.

move-asset

Move an asset to a different folder. Useful for sorting batch outputs into per-campaign folders.

share-asset

Generate a shareable link for an asset — preview-only or download.

delete-asset

Permanently delete an asset. Used by cleanup agents and retention pipelines.

Workspace & Brand Kits

6 tools

Multi-brand brand kits, folders, and credit visibility. Your agent always knows which brand it's working on and how much budget is left.

get-current-workspace

Returns the active workspace — plan tier, member count, default folder ID.

get-brand-kit

Fetch a brand kit's colors, typography, logos, and brand voice. The same fields the Playcut studio uses to shape on-brand generation.

list-brand-kits

List every brand kit in the workspace. Agency accounts can run dozens — one per client brand.

get-credit-balance

Read total + subscription + purchased credit balance plus current-period usage. Use this before queueing a batch.

create-folder

Create a new folder for organizing generations — per campaign, per client, per SKU.

list-folders

List existing folders in the workspace.

Sessions & Tasks

5 tools

Every generation runs inside a session and reports progress via task IDs. Block-until-done, poll, or fan out parallel sessions for batch work.

create-session

Create a new session — STUDIO, BOARD, or ACTOR_SESSION — to thread related generations together.

list-sessions

Paginated session listing with metadata.

get-task-status

Get current status + progress + output for any generation task.

list-tasks

List the workspace's recent tasks. Useful for agents that orchestrate long-running batches.

wait-for-task

Block until a task reaches a terminal state (COMPLETED / FAILED / CANCELLED). Polls under the hood — your agent stays clean.

Batch & Automation

1 tool

Batch operations make 50+ parallel variants feel like a single call. Built for agentic workflows.

generate-batch

Fan out a batch of generations in a single call — multiple prompts, languages, or aspect ratios at once. The orchestration is server-side.

Real workflows

What teams are automating today.

Slack → 50 UGC ad variants

01/04

A marketer types a request in Slack. A Claude agent reads the brief, finds the active actor, finds the brand kit, fans out 50 actor-video calls across 5 languages, and posts the rendered MP4s back in Slack with the credit cost. Brief to live: ~14 minutes.

// Claude prompt
"Generate 50 UGC ad variants of @mira holding the new
serum, 5 languages × 10 variants each, ad-native 9:16"

Shopify CSV → e-com catalog

02/04

A merchandiser drops a Shopify product CSV into a Cursor workspace. An agent reads each SKU, runs generate-image-from-reference + actor-shoot in parallel, and writes 12 brand-kit-consistent images per product back to a /catalog folder ready for Shopify upload.

// Cursor agent script
products.forEach(p =>
  playcut.generateImageFromReference({
    prompt: p.scene,
    referenceImages: [p.uploadId],
    aspectRatio: "1:1",
  })
)

Abandoned-cart flow → personalized recovery ads

03/04

An n8n workflow fires on cart-abandonment events. It calls Playcut to render a personalized 15-second video featuring the brand's custom actor holding the specific abandoned product, then ships the MP4 to Klaviyo for the recovery email. Per-recipient creative, automated.

// n8n node
POST https://api.playcut.ai/v1/actor-video
{
  actorId: "mira_chen_001",
  prompt: "Holding {{cart.product.name}} in soft window light",
  duration: 15
}

Agentic UGC pipeline

04/04

A long-running Claude agent runs on a cron — every Monday it audits last week's ad performance from Meta + TikTok, identifies winners, generates 20 fresh variants of the top creative via Playcut, ships them back into Meta Ad Manager with naming conventions for tracking, and reports the cost in a Notion log.

// Claude agent (excerpt)
const winners = await meta.topAds({ days: 7 });
const variants = await playcut.generateBatch({
  baseActor: winners[0].actorId,
  count: 20,
  ...
});

Authentication

API keys. Bearer tokens.

Generate keys in your workspace settings. Scoped to your workspace, revocable at any time, never exposed in MCP responses. All traffic is TLS. Rotate quarterly, one key per host machine or agent runtime.

Bearer-token auth · TLS-only
Per-workspace scoping · multi-tenant safe
Revocable + rotatable from the UI
Concurrency cap: 3 active jobs (raisable)

Pricing

Free during open beta.

MCP + API access is bundled into every Playcut plan at no extra charge. You only pay for the credits each generation costs — the same pricing as if you'd run it in the app.

01 Hobby $9/mo · 500 credits · 3 custom actors
02 Pro $29/mo · 2,000 credits · 10 actors
03 Studio $79/mo · 6,000 credits · 25 actors
✓ Agency $149/seat/mo · 10,000 credits · unlimited actors
05 All plans · API + MCP included

Questions

MCP & API FAQ

What is the Playcut MCP server? +

Playcut's MCP server exposes the entire Playcut studio as Model Context Protocol tools. Any MCP-compatible AI client — Claude Desktop, Claude Code, Cursor, Windsurf, Zed, or a custom agent — can call build-actor, image-gen, video-gen, voice-clone, brand-kit, batch-generate, and dozens more operations the same way you'd click around the app. 37 tools total, covering every workflow surface.

How do I install the Playcut MCP server? +

Generate an API key in your Playcut workspace settings, then add the Playcut MCP entry to your client's config — claude_desktop_config.json for Claude Desktop, .mcp.json for Claude Code, mcp.json for Cursor. The server is hosted, so there's no local binary to install. Setup takes under 60 seconds. Full install guides are linked from your workspace settings page.

Is there a REST API too? +

Yes. Every MCP tool has a 1:1 REST equivalent at api.playcut.ai. The MCP server is a thin wrapper over the same API — use whichever fits your stack. n8n, Zapier, Make, custom backends, and serverless functions all work with the REST surface. API keys are the same.

Which AI clients work with Playcut MCP? +

Any client that speaks the Model Context Protocol. Today that includes Claude Desktop (1-click .mcpb install), Claude Code (CLI), the Claude API (direct tool-use), Cursor, Windsurf, Zed, and any custom MCP host. For non-MCP clients like ChatGPT custom GPTs and Zapier/n8n/Make, use the REST API.

Can I generate AI actors programmatically? +

Yes — that's the headline use case. generate-actor creates a custom AI actor from a structured description (sex, age, ethnicity, hair, body, signature accessories), returns an actor ID, and that ID is reusable forever. actor-shoot and actor-video then render stills and motion video featuring that same actor with 100% character consistency. No other AI video API offers this depth.

What does pricing look like for API + MCP usage? +

MCP + API access is free during the open beta. All Playcut plans (Hobby $9, Pro $29, Studio $79, Agency $149/seat) include API access at no extra charge — you only pay for the credits each generation costs, the same as if you'd run it in the app. Credit pricing matches the app: typical image ~5 credits, typical actor shoot ~20 credits, typical short video ~40+ credits.

How is auth handled? Is it secure? +

Bearer-token auth via API keys you generate in your workspace settings. Keys are scoped to your workspace, revocable at any time, and never exposed in MCP responses. All traffic is TLS. Recommended pattern: one API key per host machine / agent runtime, rotated quarterly.

Are there rate limits? +

Per-workspace concurrency cap of 3 active generation jobs by default, raisable on Studio / Agency plans. REST and MCP share the same quota. wait-for-task and get-task-status are unmetered. There is no API request-per-second limit beyond the active-job concurrency cap.

Build the AI video workflow
your competitors can't.

Get an API key, drop one config block into your editor, and start automating today. Every Playcut tool — actors, video, voice, brand kits, batches — at your agent's fingertips.

Get your API key Browse the 37 tools

The complete AI video studio, in your AI editor.