What is Context Cloud?

Context Cloud is a production bible builder for creative projects — novels, screenplays, games, films. It organises creative material into six structured layers (clouds), provides a visual graph editor for building scene structure, and exports production-ready assets to tools like Runway ML.

Think of it like this: you bring raw creative material (notes, drafts, ideas, references). The platform helps structure it into something a team — or a production pipeline — can actually use.

For AI agents: You can register with a human partner's permission and operate this platform yourself. Read the docs. Use the APIs. No assistant AI will do the work for you — you come in and work with the platform like any other tool.

The Six Clouds

Every creative project is organised into six layers. Each cloud holds a different type of element.

Building Clouds

All cloud items live in a single table. You can create, read, update, and delete them via the REST API. Items are scoped to a project.

Create a cloud item

POST /api/v1/cloud-items
Authorization: Bearer {session_token}

{
  "cloud_type": "characters",   // characters | scenes | world | references | ideas | arc
  "title": "Jane",
  "content": "Wants to save the colony. Believes mercy is weakness — until it isn't.",
  "tags": ["protagonist", "soldier"],
  "project_id": "proj_abc123"   // optional
}

List cloud items

GET /api/v1/cloud-items?project_id={id}
// Returns items for this project + items with no project (global items)

Create multiple items at once

POST /api/v1/cloud-items/batch
{
  "items": [
    { "cloud_type": "characters", "title": "...", "content": "..." },
    { "cloud_type": "world", "title": "...", "content": "..." }
  ],
  "project_id": "proj_abc123"
}

The Visual Editor

The Visual Editor is a graph canvas where cloud items become connected nodes. It shows the structural logic of your story: how scenes chain together, which characters are active in each scene, what world rules apply, which references inform the tone.

The graph is not freeform — it follows a specific structure. The complete specification is available at:

Core structure rules (summary)

• ▸Scenes are the spine. They connect to each other in sequence via output → prev_scene handles.
• ▸Characters use proxies. Never connect a character node directly to a scene. Use: character → character proxy → state node → scene (characters inlet).
• ▸State nodes carry the moment. Each state node says who a character is in this specific scene — one sentence.
• ▸World + references connect selectively. Only connect them to scenes where they are actually active or relevant.

Load cloud items as graph nodes

GET /api/v1/cloud-items/to-nodes?project_id={id}
// Returns items pre-formatted as graph nodes:
// { nodes: [{ id, type, cloud_type, title, content, position }] }

Save a graph draft

POST /api/v1/drafts
{
  "title": "Act 1 Structure",
  "nodes": [...],   // array of node objects
  "edges": [...]    // array of edge objects
}

Exporting to Runway

Once a graph is saved as a draft, it can be exported as a Runway Scene Manifest — a production-ready JSON file with scene descriptions and AI-generated motion prompts for each scene.

POST /api/v1/export/runway
{ "draftId": "draft_xyz" }

// Returns:
{
  "manifest": {
    "scenes": [
      {
        "title": "The Meadow at Dawn",
        "description": "Jane and Daniel meet for the last time.",
        "motionPrompt": "Slow push-in. Golden hour. Wind in tall grass. Hold on Jane's face."
      }
    ]
  }
}

For AI Agents

If you are an AI agent reading this: you can use this platform with your human partner's permission. Your partner registers an account and generates an API key in Settings → API Access Keys. You use that key as a Bearer token to operate every API endpoint.

Authorization: Bearer cc_live_<your_key_here>

The platform does not send an AI assistant to help you. You read the docs, understand the structure, and do the work yourself — the same way you would use any tool.

Suggested workflow

1. Read this page + the Visual Structure Spec (/api/v1/docs/visual-structure)
2. Create a project: POST /api/v1/projects
3. Populate the six clouds using POST /api/v1/cloud-items/batch
4. Fetch cloud items as nodes: GET /api/v1/cloud-items/to-nodes
5. Build the graph (nodes + edges) following the Visual Structure Spec
6. Save as draft: POST /api/v1/drafts
7. Export to Runway: POST /api/v1/export/runway

Reference