New: One-click OAuth connect

The ultimate video creation MCP for AI agents

Give Claude, ChatGPT, Cursor, internal automations, and custom AI agents a direct way to create viral-ready videos with Revid. Connect the MCP with one-click OAuth sign-in, script it with the CLI, or call the public API directly when you want full backend control.

Works withClaudeClaude CodeChatGPTCursorWindsurfVS CodeAny MCP client

One-click OAuth connect

Paste the server URL in Claude, ChatGPT, Cursor, or VS Code and sign in with your Revid account. No API key to copy, scoped tokens with automatic refresh.

Generate finished videos

Turn scripts, prompts, articles, audio, music, captions, avatars, and ads into short-form videos.

Built for autonomous agents

Expose the full Revid creation pipeline through MCP tools with clear resources and structured responses.

One endpoint, many workflows

Use the MCP wrapper for agent actions or call the same public API directly from your own backend.

MCP, CLI, or API? Pick your interface

All three wrap the same Revid video creation pipeline. Choose based on where your agent or code runs.

Revid MCP

Tool-calling agents

Claude, ChatGPT, Cursor, VS Code, and custom agents built on the MCP SDKs. Connect with one-click OAuth or an API key.

https://www.revid.ai/api/mcp

Revid CLI

Terminal agents, CI jobs, scripts

The revid-cli npm package with stable --json output, exact API payload mode, render polling, exports, and credit estimates.

npm install -g revid-cli

Public API

Backend integrations and products

Call the same video creation API directly from your own code when you want full control over rendering and publishing.

POST https://www.revid.ai/api/public/v3/render
Quick connect

Connect your agent in under a minute

Hosted MCP clients connect with OAuth: paste the server URL, sign in with your Revid account, done. Header-configurable clients can use an API key instead.

OAuth 2.1, no API key needed
  1. 1Open Claude settings and go to Connectors.
  2. 2Click "Add custom connector" and paste the Revid MCP URL.
  3. 3Click Connect, sign in with your Revid account, and approve access.
Remote MCP server URL
https://www.revid.ai/api/mcp

Your API key

Checking your session so we can personalize the examples with your Revid API key.

Keep this key secret. Requests made with it, or with OAuth tokens, consume credits from your Revid account or active workspace.

OAuth under the hood
# Revid MCP OAuth 2.1 details
flow: authorization_code + PKCE (S256)
scope: mcp:use
dynamic client registration: POST https://www.revid.ai/api/oauth/register
authorization endpoint: https://www.revid.ai/api/oauth/authorize
token endpoint: https://www.revid.ai/api/oauth/token
protected resource metadata: https://www.revid.ai/.well-known/oauth-protected-resource
authorization server metadata: https://www.revid.ai/.well-known/oauth-authorization-server
access tokens: 1 hour; refresh tokens: rotating, long-lived
fallback auth: API key via Authorization Bearer, x-revid-api-key, x-api-key, or key header
Direct API render call
curl -X POST "https://www.revid.ai/api/public/v3/render" \
  -H "Content-Type: application/json" \
  -H "key: YOUR_REVID_API_KEY" \
  -d '{
    "workflow": "voiceover-to-video",
    "source": {
      "text": "Create a 30-second short video about 3 surprising AI productivity tips."
    },
    "media": {
      "type": "moving-image",
      "quality": "pro",
      "animation": "soft"
    },
    "voice": {
      "enabled": true,
      "voiceId": "cgSgspJ2msm6clMCkdW9"
    },
    "captions": {
      "enabled": true,
      "preset": "Wrap 1",
      "position": "bottom"
    }
  }'
Agent-readable reference
# Revid MCP quick reference
endpoint: https://www.revid.ai/api/mcp
transport: Streamable HTTP
auth (hosted clients): OAuth 2.1 with PKCE and dynamic client registration; discovery via https://www.revid.ai/.well-known/oauth-protected-resource
auth (header clients): Authorization Bearer, x-revid-api-key, x-api-key, or key header with a Revid API key
primary flow: render_video -> get_project_status -> export_video or publish_now
render docs resource: revid://docs/render
mcp guide resource: revid://docs/mcp
agent-readable docs: https://www.revid.ai/llms.txt, https://www.revid.ai/llm/revid-ai-mcp-server.md
tools: generate_media, render_video, get_project_status, list_projects, calculate_credits, export_video, rename_project, clone_voice, list_characters, create_character, delete_character, schedule_publish, publish_now, buy_credit_pack
Agent workflow

From prompt to published video

The MCP wraps Revid's public video API, so agents can estimate, generate, monitor, export, rename, and publish projects without inventing their own video pipeline. Agents can read the `revid://docs/render` resource for the always-current payload reference.

1

Optionally call `calculate_credits` to estimate cost before rendering.

2

Call `render_video` with a workflow and source payload, then store the returned `pid`.

3

Poll `get_project_status` until the response includes `videoUrl`.

4

Optionally call `export_video`, `schedule_publish`, or `publish_now`.

Supported video workflows

Agents can choose the workflow that matches the user's intent, then pass only the relevant source, media, voice, caption, music, avatar, and render fields.

voiceover-to-videoscript-to-videoprompt-to-videoarticle-to-videoaudio-to-videomusic-to-videoavatar-to-videocaption-videoad-generator

MCP tools exposed

The tool names are stable and descriptive, which makes this page easy for humans to scan and easy for AI agents to parse.

generate_media
render_video
get_project_status
list_projects
calculate_credits
export_video
rename_project
clone_voice
list_characters
create_character
delete_character
schedule_publish
publish_now
buy_credit_pack

Build video agents that actually ship finished content

Revid handles scripting, visuals, voice, captions, music, rendering, exporting, and social publishing, so your agent can focus on intent, strategy, and iteration.

Revid CLI

Create Revid videos from any terminal agent

The `revid-cli` package is the fastest way to give shell-based agents, CI jobs, and local scripts access to Revid video creation. It supports exact API payloads, stable `--json` output, guided human setup, async waiting, exports, credits, docs, and project listing.

Install
npm install -g revid-cli
export REVID_API_KEY="your_revid_api_key"
revid --help
Agent commands
revid create --prompt "A cinematic 30 second product launch video" --wait --json
revid create --payload payload.json --wait --json
revid projects --limit 10 --json
revid status <pid> --json
revid export <pid> --json
MCP workflows in action

Show agents what Revid can create

These examples map directly to Revid MCP and API workflows, from prompts and PDFs to motion transfer and music videos. Use them as a visual reference for what an agent can generate after calling `render_video`.

AI Magic

Turn simple ideas into
incredible videos

See how Revid transforms basic text, images, or documents into studio-quality videos in minutes.

YOUR IDEA
Tool:AI Movie Maker
Your Prompt

Script automatically generated by Revid

SCENE 1 – THE STRANGE BEGINNING Exterior, Modena city walls, early morning, 1325. Dark clouds gather over medieval Italy. The camera slowly pushes in toward MATTEO OF MODENA, a battle-worn soldier in steel armor and chainmail, standing waist-up on the stone wall facing the camera. Soldiers move behind him preparing weapons. [Matteo of Modena – Veteran Soldier of Modena] (looking directly at the viewer) You’d think wars start over land… or gold. But the war I fought in… started because of a bucket. --- SCENE 2 – THE BUCKET Camera slowly circles Matteo as he walks along the city well in Modena’s courtyard. He bends down and lifts an old wooden bucket. Townspeople move nervously in the background. [Matteo] A few of our men snuck into Bologna one night. They didn’t steal treasure… or weapons. (raises the bucket slightly toward camera) They stole this. A simple bucket from the city well. --- SCENE 3 – THE INSULT Camera pans with Matteo as he walks toward the edge of the wall overlooking distant hills. He points across the landscape. The city of Bologna can be seen faintly in the distance. [Matteo] Over those hills… lies Bologna. And to them… this wasn’t a joke. It was an insult. --- SCENE 4 – AN ARMY FOR A BUCKET Wide shot as Matteo walks through soldiers preparing armor and sharpening swords. The camera tracks backward while he walks toward it. Behind him thousands of soldiers gather. [Matteo] So Bologna did something unbelievable. They raised an army. Thirty thousand men… All marching here… For a bucket. --- SCENE 5 – THE BATTLE OF ZAPPOLINO Camera cuts to Matteo standing waist-up on a dusty battlefield. Smoke rises. Soldiers clash behind him. The camera slowly pushes in while arrows fly overhead. [Matteo] We had barely seven thousand. But we met them at Zappolino. Steel crashed. Arrows filled the sky. And suddenly… A joke turned into a war. --- SCENE 6 – CHAOS OF WAR Camera handheld now, moving slightly to create urgency. Matteo ducks slightly as soldiers rush past him fighting behind him. He gestures toward the chaos behind him. [Matteo] Look behind me. That’s Bologna’s army. Thirty thousand men… Running. --- SCENE 7 – THE RETREAT Tracking shot as Matteo walks forward through the battlefield while defeated soldiers flee behind him. [Matteo] The great army of Bologna broke. And they ran all the way back to their city walls. We chased them the entire way. --- SCENE 8 – THE TROPHY Camera slows and centers Matteo again. He now stands inside a stone tower room lit by torches. The bucket sits on a pedestal beside him. He places a hand on it. [Matteo] When the fighting ended… We still had the bucket. --- SCENE 9 – THE LEGEND The camera slowly pushes in close on Matteo's face. The room is quiet now. [Matteo] Seven hundred years later… People still come to see it. --- SCENE 10 – FINAL LINE Matteo lifts the bucket slightly and looks directly at the camera. [Matteo] The strangest trophy in history. A war… Fought over a bucket.
Visual FormatAI Video
Visual StyleUltra Realism
Characters
Matteo of Modena
Matteo of Modena
FINAL VIDEO

Frequently asked questions

What is the Revid MCP?

The Revid MCP is a Streamable HTTP Model Context Protocol server that lets AI agents call Revid video creation tools, including rendering, status polling, exporting, publishing, voice cloning, and character management.

How do I connect Claude, ChatGPT, or Cursor to the Revid MCP?

Add https://www.revid.ai/api/mcp as a remote MCP server or custom connector, then sign in with your Revid account when prompted. Revid supports OAuth 2.1 with PKCE and dynamic client registration, so hosted clients connect without pasting an API key.

Do I need an API key to use the Revid MCP?

Not for OAuth-capable clients like Claude, ChatGPT, Cursor, and VS Code: you just sign in with your Revid account. Header-configurable MCP clients, the Revid CLI, and direct API calls use a Revid API key from your account page.

How do AI agents authenticate with Revid?

Hosted remote MCP clients authenticate with OAuth 2.1 and PKCE, discovered automatically from the server's well-known metadata. Header-configurable clients can authenticate with a Revid API key using Authorization Bearer, x-revid-api-key, x-api-key, or key headers.

What is the Revid CLI?

The Revid CLI is the revid-cli npm package. It lets humans, scripts, CI jobs, and AI agents create videos from the terminal using stable JSON output, exact API payload mode, render polling, exporting, and project commands.

Can I use the Revid API without MCP?

Yes. The public API can be called directly from your backend. MCP wraps the same Revid video creation API for agent-native tool calling.

How much does creating a video with the MCP or API cost?

Renders consume credits from your Revid account or active workspace. Agents can call the calculate_credits tool, or the CLI credits command, with their intended payload before rendering to estimate the exact cost.