generated: '2026-09-03'
method: probed
source: https://imaginode.ai/api/mcp (live tools/list) + https://imaginode.ai/en/docs/api-mcp
name: Imaginode MCP Server
description: 'Official hosted MCP server exposing Imaginode''s generation pipeline (images, video, model
  catalog, job status) to any MCP client over Streamable HTTP. Stateless; billed against the connected
  account''s shared credit balance at the same per-generation prices as the canvas. tools/list answers
  anonymously with full inputSchemas; tool calls require an account API key (imk_ prefix) sent as "Authorization:
  Bearer imk_...". Each key is limited to 10 generations per minute and failed generations are refunded
  automatically.'
deployment:
  mode: remote
  endpoint: https://imaginode.ai/api/mcp
  auth: api-key
  verified: probed
  probe_prior: (never probed)
  probe: live
  probe_why: live
  checked: '2026-09-11'
  source: claimed-backlog re-probe 2026-09-11
transport: streamable-http
status: official
client_setup:
  claude_code: 'claude mcp add --transport http imaginode https://imaginode.ai/api/mcp --header "Authorization:
    Bearer your-key"'
  note: In Claude.ai or Cursor, add a custom connector with the same address and the same header.
tools:
- name: list_models
  title: List Imaginode models
  description: List the available AI models with their kind (image, video), default cost in credits (1
    credit = $0.012, or 0.01 EUR when billed in euros) and billable options.
  parameters:
  - kind
- name: generate_image
  title: Generate an image
  description: Generate an image from a text prompt (optionally with input images for editing or reference).
    Waits for the result and returns the image URL. Costs credits; failures are refunded.
  parameters:
  - prompt
  - model
  - image_urls
  - aspect_ratio
  required:
  - prompt
- name: generate_video
  title: Generate a video
  description: Start a video generation from a text prompt (optionally with a start image). Returns immediately
    with a request_id and the cost in credits; poll get_generation_status every 30 seconds until it completes.
  parameters:
  - prompt
  - model
  - image_url
  - duration
  - resolution
  - aspect_ratio
  - audio
  required:
  - prompt
- name: get_generation_status
  title: Get generation status
  description: Check a generation started with generate_video (or a generate_image that timed out). Returns
    running, completed (with output URLs) or failed (refunded).
  parameters:
  - request_id
  required:
  - request_id
tool_schemas: mcp/imaginode-mcp-tools.json
x-evidence:
  fetched: '2026-09-03'
  url: https://imaginode.ai/api/mcp
  request: POST tools/list (anonymous)
  http_status: 200
  content_type: application/json