How to Use Pictory API with Pictory AI MCP to Automate Video Production Using Claude Code and Codex

For the past several months, I have been testing a modern way to create video content: What if you never had to open a traditional video editor, drag clips across a heavy timeline, or manually align audio tracks ever again?

Instead, what if you could open your terminal, talk directly to Claude Code or Codex, and say: “Read my latest 800-word article, break it into eight visual scenes, match stock B-roll for each sentence, and render a high-definition 1080p video with an AI presenter avatar”—and have the finished MP4 delivered to your screen in minutes?

That is exactly what the Pictory AI MCP server enables. By connecting AI coding agents to our open-source Pictory MCP Server on GitHub, you turn the official Pictory API into native tools that your AI assistant can control with simple plain-English instructions. Instead of manually assembling video assets, you let the Pictory AI MCP execute API requests directly from your coding terminal.

In this comprehensive guide, I will walk you through the full architecture, show you how to configure Claude Code, Codex, and Cursor IDE, break down all 5 master MCP tools and their endpoints, and share the exact workflow I use to automate high-converting videos while saving ~60% on subscription costs with the verified pictory coupon code.

The Problem: Why Web Dashboards Fail Compared to the Pictory API

If you create content for YouTube, TikTok, or client campaigns, you already know where the real friction lies. The bottleneck is almost never writing the idea. The bottleneck is the mechanical busywork:

• Searching stock libraries for 15 different clips that match your sentences.
• Trimming 3-second gaps and aligning background music volume so it doesn’t drown out speech.
• Slicing long podcasts into vertical 9:16 clips and animating karaoke-style subtitles word by word.
• Waiting on local machines to render video files that consume all your RAM and CPU.

Most web dashboards attempt to solve this with browser interfaces. But dashboards still require hundreds of clicks, constant dragging, and tab switching.

Model Context Protocol (MCP) eliminates the dashboard entirely.

Because Claude Code and Codex operate directly on your local files, your AI assistant can read a markdown file on your hard drive, plan the scene pacing, call the Pictory API via the Pictory AI MCP bridge, and hand you back a finished, hosted MP4 without leaving your terminal.

If you want to see this autonomous pipeline in action, watch my complete 40-second video walkthrough below:

FastMCP Architecture: How Claude Code & Codex Connect to the Pictory API via Pictory AI MCP

The bridge between your AI assistant and the Pictory API is built on FastMCP, Anthropic’s official high-speed Python framework. The Pictory AI MCP acts as an intelligent translator between the Model Context Protocol and the underlying Pictory API endpoints.

Here is how the data flows from your natural-language request to the final cloud render:

+-------------------------------------------------------------------------+
|                  AI CODING AGENT / DEVELOPER CLIENT                     |
|           (Claude Code CLI, OpenAI Codex, Cursor IDE, Windsurf)         |
+-------------------------------------------------------------------------+
                                    |
                                    | (JSON-RPC stdio communication)
                                    v
+-------------------------------------------------------------------------+
|                       PICTORY MCP SERVER ENGINE                         |
|             (FastMCP Server running locally via Python / uvx)           |
+-------------------------------------------------------------------------+
                                    |
                                    | (Secure HTTPS REST API Calls)
                                    v
+-------------------------------------------------------------------------+
|                         PICTORY CLOUD API V2                            |
|    - Storyblocks & Getty Images Engine (Millions of Stock Assets)       |
|    - ElevenLabs & Azure Neural TTS (300+ Lifelike Voices)               |
|    - HeyGen Digital Human Avatars (Synchronized Lip-Sync)               |
|    - Cloud GPU Distributed Render Cluster (1080p / 4K MP4 Output)       |
|    - Direct Export Connectors (Amazon AWS S3 & Vimeo Showcases)         |
+-------------------------------------------------------------------------+

1. Natural Language Input: You give Claude Code a task: “Take script.md, select an energetic electronic background track, and render a 16:9 video.”
2. Local Protocol Bridge: Claude Code translates your request into a structured JSON-RPC call sent over standard input/output (stdio) to pictory-mcp.
3. Pictory API Execution: The MCP server signs the request using your PICTORY_API_KEY and dispatches it to Pictory’s cloud endpoints.
4. Cloud Rendering & Status: Pictory’s cloud cluster handles the heavy GPU rendering, matches licensed footage, and returns real-time progress percentages until the final MP4 URL is ready.

Step-by-Step Setup: How to Configure Pictory AI MCP with Your Pictory API Key

Setting up the server takes less than two minutes because the package runs natively via uvx (the ultra-fast Python package runner from Astral). You do not need to compile code or manage virtual environments manually.

Step 1: Obtain Your Pictory API Key

To connect to the API, you need an active Pictory account with API access enabled:

1. Sign in to your dashboard at app.pictory.ai.
2. Go to Settings → API Access (or visit https://app.pictory.ai/api-access).
3. Generate and copy your API Key (pictai_...).

[!TIP]
Community Discount Stack (~60% Off): If you are signing up or upgrading, you can combine Pictory’s 40% annual plan discount with verified partner coupon code ACTIVATECODE for an extra 20% recurring lifetime discount. That saves up to $94 every year. Review our step-by-step pictory coupon code checkout guide to see the exact stacking math.

Step 2: Configure Pictory AI MCP in Claude Code (CLI)

For developers using Anthropic’s official terminal agent Claude Code, add the MCP server using a single command:

claude mcp add pictory -- uvx pictory-mcp

When prompted, set your environment variable:

export PICTORY_API_KEY="pictai_your_actual_key_here"

Alternatively, you can edit your global claude.json configuration file:

{
  "mcpServers": {
    "pictory": {
      "command": "uvx",
      "args": ["pictory-mcp"],
      "env": {
        "PICTORY_API_KEY": "pictai_your_actual_key_here"
      }
    }
  }
}

Step 3: Configure Pictory AI MCP in Claude Desktop

If you prefer the visual Claude Desktop app, edit your configuration file:
– Windows: %APPDATA%\Claude\claude_desktop_config.json
– macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Add the pictory entry inside the mcpServers block:

{
  "mcpServers": {
    "pictory": {
      "command": "uvx",
      "args": ["pictory-mcp"],
      "env": {
        "PICTORY_API_KEY": "pictai_your_actual_key_here"
      }
    }
  }
}

Restart Claude Desktop. You will immediately see a hammer icon appear in the chat interface showing all 5 Pictory tools active and ready.

Step 4: Configure Pictory AI MCP in Cursor IDE & Codex

If you build applications in Cursor IDE:

1. Open Settings → Features → MCP Servers.
2. Click + Add New MCP Server.
3. Fill in the three fields:
   – Name: pictory
   – Type: command
   – Command: uvx pictory-mcp
4. In the environment variable section, add:
   PICTORY_API_KEY=pictai_your_actual_key_here

Your Cursor Composer agent can now create, inspect, and patch video projects directly alongside your application code.

Pictory AI MCP Master Reference: All 5 Tools & Endpoints Explained

The Pictory AI MCP consolidates over 40 individual Pictory API endpoints into 5 intuitive master tools. Here is the complete breakdown of every tool, action, and capability available to your AI assistant:

1. pictory_video (Pictory API Cloud Video Engine & Rendering)

This is the primary workhorse tool in the Pictory AI MCP, used to interact with the core Pictory API video rendering endpoints.

2. pictory_transcribe (Audio Extraction & Viral Shorts)

Powered by the Pictory API speech-to-text and NLP models, this Pictory AI MCP tool handles long-form transcription and viral highlight extraction.

3. pictory_assets (Stock Footage, Avatars & Voices)

Gives your coding agent direct programmatic access to the Pictory API media catalog of over 10 million licensed Storyblocks and Getty Images assets.

4. pictory_projects (Web Dashboard & Template Sync)

Ensures that every project created via code remains 100% synchronized with your official web dashboard at app.pictory.ai.

5. pictory_integrations (Cloud Storage & Quota Management)

Monitors your Pictory API subscription quotas, credits, and direct publishing channels via the Pictory AI MCP.

5 Practical Blueprints: How to Use Pictory API to Automate Video Production

To show you how these tools come together in practice, here are five real-world workflows you can run directly inside Claude Code or Codex:

Blueprint 1: Convert Blog URLs into 1080p Videos with AI Presenters

Give Claude a public article URL. The agent scrapes the page, extracts the narrative points, matches B-roll, and adds a talking-head presenter:

{
  "videoName": "Cloud_Migration_Tutorial",
  "aspectRatio": "16:9",
  "avatar": {
    "avatarId": "Brandon_expressive2_public",
    "position": "bottom-right",
    "width": "25%",
    "borderRadius": "100"
  },
  "voiceOver": {
    "enabled": true,
    "aiVoices": [{"speaker": "Martin", "speed": 100, "volume": 100}]
  },
  "backgroundMusic": {
    "enabled": true,
    "autoMusic": true,
    "volume": 0.12
  },
  "scenes": [
    {
      "blogUrl": "https://softreviewed.com/how-to-use-pictory-api-with-pictory-ai-mcp/"
    }
  ]
}

• Outcome: Your written article becomes an engaging video showcase with a professional digital presenter in under three minutes.

Blueprint 2: Repurpose Long Podcasts into 9:16 Viral Shorts via Pictory API

Upload a 45-minute MP4 file or paste a public video URL. Claude calls pictory_transcribe:

1. It transcribes the conversation with word-level accuracy.
2. It runs engagement scoring to isolate the most dramatic 45-second soundbite.
3. It crops the video to vertical 9:16 format, centers the speaker, and burns in high-contrast karaoke captions.

• Outcome: 5 to 10 ready-to-publish TikToks, Instagram Reels, and YouTube Shorts extracted from a single recording session without manual clipping.

Blueprint 3: Zero-Credit Storyboard Preview Safeguard in Pictory API

One of the biggest concerns creators have with video APIs is burning credits on bad footage. The Pictory MCP solves this through its Zero-Cost Preview Protocol:

Step 1: Claude calls pictory_video ("create_preview") -> 0 Credits Consumed
Step 2: Inspect returned scene thumbnail URLs & sentence timings
Step 3: If scene 3 has the wrong B-roll, call pictory_projects ("update_scene")
Step 4: Claude calls pictory_video ("render_from_preview") -> 1080p Cloud Render

By verifying the scene thumbnails beforehand, you never waste a single minute of your subscription quota on a video that isn’t already 100% approved.

Blueprint 4: Automate Faceless YouTube Channels via Claude Code & Pictory API

You can instruct Claude Code to automate an entire YouTube production routine:

1. Scriptwriting: Claude writes a punchy 5-minute educational script on tech news.
2. Music Curation: Claude calls pictory_assets (action="search_music", mood="Inspiring") and selects a licensed track.
3. B-Roll Selection: Claude queries pictory_assets (action="search_media") to find verified 1080p clips for each key phrase.
4. Rendering: Claude calls render_storyboard with 16:9 framing and high-contrast subtitles.

• Outcome: Complete hands-free production for technical tutorials, documentation videos, and news recaps.

Blueprint 5: Enterprise AWS S3 Storage Synchronization via Pictory API

For agencies managing multiple clients, you don’t want completed videos sitting on local desktop drives.

Using pictory_integrations (create_aws_connection), every rendered video automatically uploads to an Amazon S3 bucket with secure access keys. Your deployment scripts can immediately serve the video to client portals or trigger CDN caching automatically.

Pictory API Cost Comparison: How to Save ~60% with ACTIVATECODE

Pictory is available under both monthly and annual plans. For developers and creators building high-volume pipelines, here is the exact pricing breakdown and how you can maximize your savings:

By selecting annual billing at checkout and entering coupon code ACTIVATECODE, the 20% recurring discount applies directly to the already-discounted annual price, giving you the lowest possible entry point for programmatic AI video generation.

Pictory API & Pictory AI MCP: Frequently Asked Questions

Can I run Pictory AI MCP with local models or OpenAI Codex?

Yes. Model Context Protocol is model-agnostic. While Anthropic’s Claude Code and Claude Desktop offer native support, any client supporting JSON-RPC stdio (including Cursor running GPT-4o, Codex, or Windsurf) can interact with this server.

Does generating storyboard previews in Pictory API consume render minutes?

No. Storyboard previews, scene text editing, and thumbnail inspections consume zero credits. Monthly rendering minutes are only deducted when you trigger a final MP4 export via render_storyboard or render_from_preview.

Are projects created via Pictory AI MCP accessible in the web app?

Yes. Every project created through the MCP server generates a real project_id. You can open app.pictory.ai in your browser at any time to visually edit or preview the project alongside the API.

What video resolutions are supported by the Pictory API?

Pictory supports vertical 9:16 (Shorts/Reels/TikTok), horizontal 16:9 (YouTube/Web), and square 1:1 (LinkedIn/Instagram) in full 1080p high definition, with 4K export available on enterprise tiers.

Final Thoughts: Getting Started with Pictory API and Pictory AI MCP

Video production is undergoing the same transition that software development experienced with infrastructure-as-code. Instead of manually clicking buttons inside proprietary graphical user interfaces, you can now define your video structure programmatically and let AI coding assistants orchestrate the entire pipeline.

To start building your own autonomous pipeline today:
1. Clone or inspect the open-source code at github.com/softreviewed/pictory-coupon-code-mcp.
2. Connect the server to Claude Code via claude mcp add pictory -- uvx pictory-mcp.
3. Unlock the lifetime 20% discount on your API plan using code ACTIVATECODE via our pictory coupon code activation guide.