Resources & Prompts

Beyond tools, the Lazer MCP server exposes resources (read-only data) and prompts (workflow templates) that help AI agents understand your production.

Resources (7)

Resources provide context about the Lazer schema, project structure, and documentation. AI agents can read these to understand the data model before making tool calls.

Schema Resources

URI	Description
lazer://schema/prisma	Full Prisma schema (all models, fields, relations)
lazer://schema/enums	All enum values as JSON (asset types, statuses, etc.)
lazer://schema/status-transitions	Valid asset status transitions with the full state machine

Project Resources

URI	Description
lazer://project/{id}/tree	Complete project hierarchy (scenes → shots → assets)

The project tree resource is a template — replace {id} with a real project ID.

Documentation Resources

URI	Description
lazer://docs/product-research	Lazer UX and product research document
lazer://docs/product-plan	AI film market and execution plan
lazer://docs/system-blueprint	Production AI system architecture

Example: Reading a resource

{
  "method": "resources/read",
  "params": {
    "uri": "lazer://schema/status-transitions"
  }
}

Returns the valid transitions:

{
  "draft": ["in_review", "rejected"],
  "in_review": ["approved", "revision_requested", "rejected"],
  "revision_requested": ["draft"],
  "approved": ["final"],
  "rejected": [],
  "final": []
}

Prompts (5)

Prompts are pre-built workflow templates that guide the AI through complex production tasks. Each prompt accepts parameters and returns a structured message that the AI can execute.

Prompt	Parameters	Description
scene_breakdown	projectId, sceneId	Analyze a scene and suggest shots with camera angles, movement, and descriptions
shot_planning	projectId, sceneId, style?	Create a shot-by-shot cinematography plan for a scene
asset_review	projectId, sceneId, assetType?	Evaluate generated assets and recommend which to select
prompt_refinement	projectId, sceneId, currentPrompt, platform?	Refine a generation prompt for better results
production_summary	projectId	Get production status with actionable next steps

Example: Scene breakdown

{
  "method": "prompts/get",
  "params": {
    "name": "scene_breakdown",
    "arguments": {
      "projectId": "clx...",
      "sceneId": "clx..."
    }
  }
}

The AI receives a detailed prompt that:

1. Reads the scene's story beat, source text, and tone
2. Lists existing shots (if any)
3. Suggests new shots with camera angles and movements
4. Considers the film's visual identity

Workflow: Full Scene Production

A typical AI-assisted workflow using prompts and tools:

1. production_summary — See what needs work
2. scene_breakdown — Plan shots for a scene
3. Use shot_create tools to create the planned shots
4. prompt_refinement — Craft generation prompts
5. Use asset_create or workflow_ingest to create assets
6. asset_review — Evaluate and select the best versions
7. Use asset_select to mark selections

Next Steps

• Tools Reference — Explore all available tools
• Setup Guide — Connect your AI client