Overview
An MCP allows your AI chat to work with Melius — reading and writing to your projects, canvases, and creatives, generating images and videos, and downloading outputs without ever leaving your AI tool.Getting started
Connect to Claude Desktop
Generate an API key
In Melius, go to your team space settings and generate a new API key with no expiry. Name it something like
Claude.Connect to Claude Code
Verifying the connection
To verify the connection works, open a new chat and send:“Create a new project called Hello World in Melius”
Uploading images to Melius via MCP
To allow the agent to upload images to Melius, you need to enable network egress in your Claude organization settings.Enable the required settings
Turn on both:
- Cloud code execution and file creation
- Allow network egress
Reference
Tools and capabilities
The Melius MCP server exposes the following tools. The agent picks the right ones based on your prompt — you won’t need to call them directly. Projectsproject_list— List projects in your team.project_get— Get a project by ID.project_create— Create a new project with an empty canvas.
canvas_list— List canvases in a project.canvas_get— Get canvas metadata.canvas_content— Get all nodes and edges on a canvas.canvas_create— Create a new canvas.canvas_delete— Delete a canvas and everything on it.canvas_plan_layout— Compute positions for new nodes that avoid overlapping existing content.
node_create— Create a node (text, image, video, file, or group).node_get— Get a node’s details and generated assets.node_update— Update a node’s prompt, model, or settings.node_delete— Delete a node.node_set_text— Set the literal text content of a custom_text node.node_set_version— Select a specific generation version on a node.node_group— Group a set of nodes into a container.bulk_create_nodes— Create multiple nodes — and optionally edges — in one call.
edge_create— Connect two nodes with an edge.edge_delete— Remove an edge.bulk_create_edges— Create multiple edges in one call.
run_start— Start an AI generation on a single node.run_get— Get the status and result of a run.run_latest— Get the most recent run for every node on a canvas.run_download— Download outputs from a single run.bulk_run_start— Start runs on multiple nodes at once.bulk_run_get— Get the status of a bulk run.bulk_run_wait— Wait for a bulk run to finish.bulk_run_download— Download outputs from a bulk run.creative_download— Download all creatives from a canvas as a ZIP.
file_upload— Upload a local image and attach it to a file node.file_upload_complete— Finalize an upload after the curl step.
model_list— List available AI models, variants, and supported settings.model_defaults— See which models are used when none is specified.
comment_list— List comments on a canvas.comment_create— Create a comment thread at a position on the canvas.comment_reply— Reply to an existing comment.comment_resolve— Mark a comment as resolved.
preset_list— Discover pre-built workflow templates.preset_apply— Apply a preset to a canvas.
display_canvas— Show an interactive preview of the current canvas state.get_guide— Load a domain-specific guide (e.g.cinema,ugc-ads,product-photo).team_list— List teams your API key has access to.