NewClippie API and MCP server

Make videos from code or from Claude.

Connect Claude, ChatGPT or Cursor in a minute, or call the REST API from your own code. Same models, same credits, same results as the app.

Get an API key Read the docs

Start with AI

Let your coding agent build it. You approve every spend.

Copy one prompt into Claude Code, Codex, Cursor, Claude or ChatGPT. It points your AI at the docs, keeps your key in CLIPPIE_API_KEY and checks the cost with you before anything is spent.

Or open inClaude Code Codex Cursor Claude ChatGPT

Paste it into any AI assistant. The shortcuts open with the prompt typed in; nothing is sent until you press enter.

Read the docs Every endpoint Docs for AI (llms-full.txt)OpenAPI spec

https://api.clippie.ai/mcp

The MCP server URL. Paste it into your AI client and sign in with your Clippie account.

What you can build. Everything Clippie makes, one call away.

AI Story videos
A topic or a script in, a finished narrated video out: an AI image for every scene, a voice, captions and music. Keep it editable in Clippie or get the MP4.
create_ai_story_videowrite_scriptrender_video
Clippie Recast
Put the person in any photo into a trending clip. Pick a motion template, send the photo, get the video back.
recast_videolist_motion_templates
AI video and image models
Seedance, Veo, Genjutsu, Flux, Imagen, Seedream and Nano Banana behind one API, at the same credit prices as AI Studio.
generate_videogenerate_imageupscale
Text to speech
Clippie voices and your own cloned voices, returned as an MP3 with word timings for captions.
text_to_speechlist_voices
Captions and transcripts
Words with timings, SRT and VTT from any video or audio file, in more than 100 languages.
transcribelist_caption_styles
Everything around it
Catalogs of voices, styles, music and footage, file uploads, job history and your credit balance, so an agent can plan before it spends.
get_accountestimate_costupload_file

Connect your AI. Paste one URL, sign in, approve.

Claude, ChatGPT, Cursor and VS Code sign in with OAuth. Your assistant checks the price and asks you before anything spends credits.

Connect Clippie to your AI

Add to Cursor Add to VS Code Add to Claude Add to ChatGPT

Cursor and VS Code open with Clippie ready to add. Claude and ChatGPT open their settings with the URL copied. Every client signs in with your Clippie account.

Connected? Make your first video.

This prompt has your AI check your balance, show you the cost and wait for your yes before it creates anything.

Or open inClaude ChatGPT

Paste it into any AI assistant. The shortcuts open with the prompt typed in; nothing is sent until you press enter.

Or set it up by hand

Claude (web and desktop)

Open Customize, then Connectors.
Choose + then Add custom connector, name it Clippie and paste the server URL.
Click Connect, sign in to Clippie and allow access.

https://api.clippie.ai/mcp

Claude Code

Run this once in your terminal.
Inside Claude Code, run /mcp, pick clippie and sign in.

claude mcp add --transport http clippie https://api.clippie.ai/mcp

Cursor

Add Clippie to ~/.cursor/mcp.json, or to .cursor/mcp.json in a project.
Cursor opens a Clippie sign-in the first time it connects.

mcp.json

{
  "mcpServers": {
    "clippie": {
      "url": "https://api.clippie.ai/mcp"
    }
  }
}

VS Code

Add Clippie to .vscode/mcp.json in your workspace, or run MCP: Add Server from the command palette.
Start the server from the file and sign in to Clippie when VS Code asks.

.vscode/mcp.json

{
  "servers": {
    "clippie": {
      "type": "http",
      "url": "https://api.clippie.ai/mcp"
    }
  }
}

ChatGPT

Open Settings, then Apps.
Under Advanced settings, turn on Developer mode.
Choose Create app, paste the server URL, choose OAuth and sign in to Clippie.

https://api.clippie.ai/mcp

Clients that send a fixed header

Create a key in Settings, API. It starts with clp_pat_ and is shown once.
Send it as Authorization: Bearer clp_pat_… instead of signing in.

Claude Code

claude mcp add --transport http clippie https://api.clippie.ai/mcp \
  --header "Authorization: Bearer clp_pat_..."

Cursor mcp.json

{
  "mcpServers": {
    "clippie": {
      "url": "https://api.clippie.ai/mcp",
      "headers": { "Authorization": "Bearer clp_pat_..." }
    }
  }
}

VS Code .vscode/mcp.json

{
  "inputs": [
    { "type": "promptString", "id": "clippie-key", "description": "Clippie API key", "password": true }
  ],
  "servers": {
    "clippie": {
      "type": "http",
      "url": "https://api.clippie.ai/mcp",
      "headers": { "Authorization": "Bearer ${input:clippie-key}" }
    }
  }
}

Any client (curl)

curl https://api.clippie.ai/mcp \
  -H "Authorization: Bearer $CLIPPIE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Or call the REST API. One request makes a video.

Create a key in Settings, then send it as Authorization: Bearer clp_pat_… to https://api.clippie.ai/v1.

01
Create the video
POST /v1/videos/ai-story answers 202 with a job straight away. Send an Idempotency-Key so a retry can never charge twice.
02
Wait for the job
GET /v1/jobs/:id?wait=30 holds the request until the job changes, so you never poll in a tight loop. The finished job has the MP4 and an editor link.
03
Or get a webhook
Pass a webhookUrl and Clippie posts job.completed or job.failed to it, signed with your secret.

ai-story.sh

# 1. Create the video. Send the same Idempotency-Key if you retry,
#    so a retry returns the first job instead of charging twice.
#    Add "dryRun": true to see the exact cost without spending anything.
curl https://api.clippie.ai/v1/videos/ai-story \
  -H "Authorization: Bearer $CLIPPIE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: 5d1c0e2a-7c4f-4a51-9b1e-2f0c6f1d9a77" \
  -d '{
    "prompt": "The lost library of Alexandria",
    "aspectRatio": "9:16",
    "durationSeconds": 60,
    "webhookUrl": "https://example.com/clippie-webhook"
  }'
# => 202 { "job": { "id": "job_...", "status": "queued", ... } }

# 2. Wait for it. ?wait= holds the request open for up to 50 seconds.
curl "https://api.clippie.ai/v1/jobs/job_...?wait=30" \
  -H "Authorization: Bearer $CLIPPIE_API_KEY"
# => { "job": { "status": "completed", "output": { "videoUrl": "...", "editorUrl": "..." } } }

ai-story.mjs

import { randomUUID } from 'node:crypto';

const API = 'https://api.clippie.ai/v1';
const auth = { Authorization: `Bearer ${process.env.CLIPPIE_API_KEY}` };
const video = { prompt: 'The lost library of Alexandria', aspectRatio: '9:16', durationSeconds: 60 };

// Check the cost first. A dry run spends nothing.
const dry = await fetch(`${API}/videos/ai-story`, {
  method: 'POST',
  headers: { ...auth, 'Content-Type': 'application/json' },
  body: JSON.stringify({ ...video, dryRun: true }),
}).then((res) => res.json());
console.log(dry.estimate); // { credits, minutes, wouldSucceed, reason }

// Make one key per video and reuse it on retries, so you are never charged twice.
const idempotencyKey = randomUUID();
let { job } = await fetch(`${API}/videos/ai-story`, {
  method: 'POST',
  headers: { ...auth, 'Content-Type': 'application/json', 'Idempotency-Key': idempotencyKey },
  body: JSON.stringify(video),
}).then((res) => res.json());

// Long-poll: each call waits up to 30 seconds for the job to change.
while (!['completed', 'failed', 'canceled'].includes(job.status)) {
  ({ job } = await fetch(`${API}/jobs/${job.id}?wait=30`, { headers: auth }).then((res) => res.json()));
}

console.log(job.status, job.output?.videoUrl, job.output?.editorUrl);

ai_story.py

import os, uuid, requests

API = "https://api.clippie.ai/v1"
auth = {"Authorization": f"Bearer {os.environ['CLIPPIE_API_KEY']}"}
video = {"prompt": "The lost library of Alexandria", "aspectRatio": "9:16", "durationSeconds": 60}

# Check the cost first. A dry run spends nothing.
dry = requests.post(f"{API}/videos/ai-story", json={**video, "dryRun": True}, headers=auth).json()
print(dry["estimate"])

# Make one key per video and reuse it on retries, so you are never charged twice.
idempotency_key = str(uuid.uuid4())
job = requests.post(
    f"{API}/videos/ai-story",
    json=video,
    headers={**auth, "Idempotency-Key": idempotency_key},
).json()["job"]

# Long-poll: each call waits up to 30 seconds for the job to change.
while job["status"] not in ("completed", "failed", "canceled"):
    job = requests.get(f"{API}/jobs/{job['id']}", params={"wait": 30}, headers=auth, timeout=60).json()["job"]

print(job["status"], (job.get("output") or {}).get("videoUrl"))

Verify the webhook. Before you trust the body.

Your signing secret (whsec_…) is in Settings, API. Check Clippie-Signature against the raw body, reject anything older than five minutes, and dedupe on Clippie-Job-Id plus the event, since a delivery can arrive twice.

verify.mjs

import { createHmac, timingSafeEqual } from 'node:crypto';

// header: the Clippie-Signature header, "t=<unix>,v1=<hex>"
// rawBody: the request body exactly as received, before JSON parsing
// secret: your whsec_... signing secret from Settings > API
export function verifyClippieWebhook(rawBody, header, secret, toleranceSeconds = 300) {
  const parts = Object.fromEntries(header.split(',').map((part) => part.trim().split('=')));
  const timestamp = Number(parts.t);
  if (!Number.isInteger(timestamp) || !parts.v1) return false;
  if (Math.abs(Date.now() / 1000 - timestamp) > toleranceSeconds) return false;
  const expected = createHmac('sha256', secret).update(`${timestamp}.${rawBody}`).digest();
  const given = Buffer.from(parts.v1, 'hex');
  return given.length === expected.length && timingSafeEqual(given, expected);
}

verify.py

import hashlib, hmac, time

# header: the Clippie-Signature header, "t=<unix>,v1=<hex>"
# raw_body: the request body bytes exactly as received
# secret: your whsec_... signing secret from Settings > API
def verify_clippie_webhook(raw_body: bytes, header: str, secret: str, tolerance: int = 300) -> bool:
    parts = dict(part.strip().split("=", 1) for part in header.split(",") if "=" in part)
    try:
        timestamp = int(parts.get("t", ""))
    except ValueError:
        return False
    if abs(time.time() - timestamp) > tolerance:
        return False
    signed = f"{timestamp}.".encode() + raw_body
    expected = hmac.new(secret.encode(), signed, hashlib.sha256).hexdigest()
    return hmac.compare_digest(expected, parts.get("v1", ""))

See the cost first
Add "dryRun": true to any create call, or call POST /v1/estimate. You get the exact credits and minutes, and nothing is spent.
Retries are safe
Replaying an Idempotency-Key within 24 hours returns the first job. The same key with a different body is a 409.
Signed webhooks
Every delivery carries Clippie-Signature: t=…,v1=…, an HMAC-SHA256 of the timestamp and body, retried for about a day.
Typed errors and limits
402 insufficient_credits, 403 subscription_required, 429 with Retry-After. 120 requests a minute and 5 running jobs per account.

Every tool, one registry. The same actions over MCP and REST.

Tools that spend credits or minutes say so, and run the same checks and prices as the app. The full request and response schemas are in the OpenAPI spec.

API reference OpenAPI spec

Create

generate_imageSpends credits
Generates 1-4 images with an AI Studio model (Nano Banana, Seedream, Flux, Imagen). Spends credits: about 6 credits for one 1-megapixel Nano Banana image; bigger sizes, more images and premium models cost more (list_models has every price). Run with dryRun: true (or estimate_cost) first and confirm the cost. Returns a job; get_job with waitSeconds returns the image URLs, usually within a minute.
POST /v1/ai/images
generate_videoSpends credits
Generates a short AI video clip with an AI Studio model: Seedance and Veo from text or images, Clippie Recast and Genjutsu from a source video (referenceVideoUrl or motionTemplateId) plus photos. Spends credits, from tens to hundreds a clip depending on model, length and resolution (list_models has the prices). Run with dryRun: true (or estimate_cost) first and confirm the cost. Returns a job; clips take 1-10 minutes, so use get_job with waitSeconds.
POST /v1/ai/videos
recast_videoSpends credits
Clippie Recast: puts the people in your photos into a trending motion template (list_motion_templates), keeping its moves, camera and sound. Send one photo per person in the template. Spends credits by the template's length: about 413 credits for a 10-second clip at 720p; 1080p costs more. Run with dryRun: true (or estimate_cost) first and confirm the cost. Returns a job; recasts take a few minutes, so use get_job with waitSeconds.
POST /v1/ai/recast
upscaleSpends credits
Upscales an image (2x-6x) or a video (to 720p, 1080p or 4K) with Topaz. Spends credits: about 7 for a 2x image (more at 4x and 6x) and 155 for a video of up to 30 seconds to 1080p at 60 fps. The price follows the real file: bigger images, longer videos (up to 10 minutes) and 4K cost more. Run with dryRun: true (or estimate_cost) first and confirm the cost. Images finish within the call; videos return a running job, so use get_job with waitSeconds.
POST /v1/ai/upscale
text_to_speechSpends credits
Turns text into an mp3 narration with a Clippie voice or one of your cloned voices, with word timings. Spends voiceover minutes for the audio's length (about 150 words a minute), and the same in caption minutes for the word timings. Finishes within the call: the job comes back completed with audioUrl and captions.words.
POST /v1/ai/speech
transcribeSpends credits
Transcribes an audio or video file (up to 30 minutes) into punctuated text, word timings, and SRT and WebVTT captions split at sentence ends. Spends caption minutes for the file's length. Finishes within the call: the job comes back completed with captions.words, captions.srt and captions.vtt.
POST /v1/ai/transcribe
create_ai_story_videoSpends credits
Makes a narrated AI Story video (script, voiceover, scene images, captions, music) and renders it to MP4. Give one of prompt (a script is written to fill durationSeconds, 30-300, default 60), script, or voiceoverFileId. Spends minutes, not credits: about 1 voiceover, 1 caption and 1 export minute per minute of video (render: false skips the export minutes and returns an editable project). Run estimate_cost (or dryRun: true) first and confirm the cost. Takes a few minutes: wait with get_job (waitSeconds). At most 3 videos are made at once per account. The result has videoUrl and editorUrl.
POST /v1/videos/ai-story
render_videoSpends credits
Renders an AI Story project you own to MP4: one made with create_ai_story_video (render: false) or in the Clippie editor. Spends export minutes for the video length (1 per minute of video). Run estimate_cost (or dryRun: true) first. Projects with placeholder scene images are refused: regenerate those scenes in the editor first.
POST /v1/videos/:projectId/render
cancel_jobCreate
Cancels a job that is still queued and refunds what it was charged. Jobs that have started cannot be canceled.
POST /v1/jobs/:id/cancel
upload_fileCreate
Imports an image, video or audio file from a public https URL (up to 100 MB) so other tools can use it as fileId. REST callers can also send the file itself as multipart/form-data (field "file").
POST /v1/files
write_scriptCreate
Writes a narration script for a short video from a topic, with the AI Story editor's writer, a theme, a tone and a target length. Free: spends no credits or minutes. Finishes within the call: the job comes back completed with output.script. Pass the script on to text_to_speech or create_ai_story_video.
POST /v1/ai/scripts

Read

get_accountRead
Shows the user's Clippie plan, AI credit balance (plan and top-up), the voiceover, caption and export minutes left, when the period ends and whether videos are watermarked. Call this first, and before anything that spends.
GET /v1/me
list_credit_historyRead
Lists credit movements (top-ups, plan credits, generations, refunds) and minute spends, newest first, each linked to its API job when there is one.
GET /v1/credits/history
list_voicesRead
Lists narration voices: the user's own cloned voices first, then the popular voices, then the rest. Filter with search (name, accent, style), gender, language or mine. Pass a voice's id as voiceId.
GET /v1/voices
list_caption_stylesRead
Lists the burned-in caption styles (font, colours, one word at a time or lines). Pass an id as captionStyleId.
GET /v1/caption-styles
list_art_stylesRead
Lists the looks an AI Story video can have. Either pass a type "style" id as artStyleId, or pass one of imageModels as imageModel together with a type "stylePrompt" id as stylePromptId. Direct-model scene images are included on paid plans.
GET /v1/art-styles
list_story_templatesRead
Lists ready-made AI Story formats (an art style plus a frame shape, e.g. "History documentary" 16:9). Pass an id as storyTemplateId to set the look and aspect ratio in one go.
GET /v1/story-templates
list_motion_templatesRead
Lists the trending motion clips used by Clippie Recast and motion-control video models, in the same order as AI Studio (newest hand-picked uploads first, then by trend). Pass an id as motionTemplateId. Each template says its length and whether Recast accepts it (recastCompatible); offer only those for recast_video.
GET /v1/motion-templates
list_modelsRead
Lists the AI Studio video and image models with their resolutions, aspect ratios, durations and credit price per generation (images: per image). Use an id as model for generate_video / generate_image.
GET /v1/models
list_background_videosRead
Lists the gameplay and satisfying background clips (Minecraft parkour, Roblox, ...) used behind story videos.
GET /v1/background-videos
list_musicRead
Lists background music tracks, popular first. Tracks with hasCopyrightWarning may be muted on some platforms. Pass an id as musicId.
GET /v1/music
list_sound_effectsRead
Lists the short sound effects (whoosh, bell, punch, ...) that can be placed on a video timeline.
GET /v1/sound-effects
list_languagesRead
Lists the languages for captions, transcription and narration. Pass a code as language.
GET /v1/languages
list_jobsRead
Lists the jobs started through the API or MCP, newest first, with their status, output and what they cost. Filter by status or kind.
GET /v1/jobs
get_jobRead
Gets one job: status (queued, processing, rendering, completed, failed, canceled), step, progress, output (videoUrl, imageUrls, audioUrl, captions, editorUrl) and cost. Pass waitSeconds (up to 50) to wait for it to finish instead of polling in a loop.
GET /v1/jobs/:id
estimate_costRead
Shows what a create tool would charge (credits and voiceover, caption and export minutes) and whether it would go through, without spending anything. Pass the tool name and the same input you would send it. Confirm the cost with the user before creating.
POST /v1/estimate
get_fileRead
Gets a file uploaded with upload_file: its type and URL.
GET /v1/files/:id

Requires a paid plan

Included in every paid plan. No API surcharge.

API keys and connected apps need an active subscription. Everything you make comes out of the same credits and minutes as the app, at the same prices, and lands in your Clippie library.

See pricing Open Settings, API

Over 5 billion views generated by Clippie. Yours could be next.

Get an API key

Make videos from code or from Claude.

Let your coding agent build it. You approve every spend.

What you can build. Everything Clippie makes, one call away.

AI Story videos

Clippie Recast

AI video and image models

Text to speech

Captions and transcripts

Everything around it

Connect your AI. Paste one URL, sign in, approve.

Connect Clippie to your AI

Claude (web and desktop)

Claude Code

Cursor

VS Code

ChatGPT

Clients that send a fixed header

Or call the REST API. One request makes a video.

Create the video

Wait for the job

Or get a webhook

Verify the webhook. Before you trust the body.

See the cost first

Retries are safe

Signed webhooks

Typed errors and limits

Every tool, one registry. The same actions over MCP and REST.

Included in every paid plan. No API surcharge.

Over 5 billion views generated by Clippie. Yours could be next.