Framedeck

What is Framedeck?

Framedeck is an MCP server that connects any AI assistant (Claude, ChatGPT, Gemini, Copilot, etc.) to the Framedeck content production pipeline. It is designed for content creators to manage productions, stages, frames, and tasks through natural language.

How to use Framedeck?

Get an API key from framedeck.app (Settings > API Keys), then configure your MCP client with the server command npx framedeck-mcp and the environment variable FRAMEDECK_API_KEY. Once configured, you can use natural language prompts to manage boards, cards, and tasks.

Key features of Framedeck

• Connect any MCP-compatible AI assistant to Framedeck.
• 32 tools for full content production management.
• Dual mode: Creator mode for content pipelines, Classic for project management.
• Idea Pool with automatic idea capture and graduation.
• Safety annotations (readOnlyHint, destructiveHint) for safe AI interactions.
• Batch create, assign, log work, and link commits.

Use cases of Framedeck

• Capture new video ideas and graduate them to full productions.
• Plan a production day by adding and tracking subtasks.
• Get a project overview with overdue items and assigned cards.
• Manage content pipeline from ideation to publishing.

FAQ from Framedeck

How do I get an API key for Framedeck?

Visit framedeck.app, go to Settings > API Keys, and generate an API key.

Which AI assistants does Framedeck work with?

It works with Claude, ChatGPT, Gemini, Copilot, and any MCP-compatible client.

What are the two modes in Framedeck?

Creator mode for content pipelines (Productions, Stages, Frames) and Classic mode for standard project management (Projects, Stages, Cards). The MCP server automatically uses the correct terminology based on each board’s mode.

What happens when I add an idea without specifying a board?

Ideas automatically land in the Idea Pool. You can then use the graduate_to_production tool to promote an idea to a dedicated production with full stages (Idea → Scripting → Filming → Editing → Published).

How are destructive actions handled?

Each tool includes MCP safety annotations (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) so AI clients can warn users before destructive actions like delete_card, delete_column, and archive_card.