midjourney-image
Generate and edit Midjourney images or videos through Ace Data Cloud. Use for imagine, edit, blend, variations, upscale, describe, video generation, and querying the exact task ID.
Install
npx skills add https://github.com/AceDataCloud/Skills/tree/main/skills/midjourney-image
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install acedatacloud-skills@llmmart
git clone https://github.com/AceDataCloud/Skills.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole acedatacloud/skills collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
Midjourney integration status
Midjourney's service and Imagine document are currently marked private. Do not submit legacy generation examples or configure mcp-midjourney as a public capability. An explicit API-spec lookup can still return a definition; that is not a publication or access grant. Check the current public catalog and let the user choose a published image service before changing their workflow; do not silently substitute a model or service.
Diagnose an existing imagine request
- Inspect the original payload locally without printing its API key.
POST /midjourney/imagineselects a model version withversion, notmodel. Remove themodelkey entirely, including when its value is an empty string ornull. - A request containing
modelreturns HTTP 400bad_requestbefore generation or asynchronous acceptance, without triggering generation or charging. For version selection the field is"version": "8.2"; omittingversionkeeps the default version. Correcting a payload does not authorize submitting it while the capability is private. - If an earlier authorized request returned a
task_id, retain that exact ID. An accepted task is not a finished image. Only report success from the actual finalresponse.success: trueand usableresponse.image_urlorresponse.raw_image_url; do not invent an artifact or treat a task ID as one. - A network disconnect, proxy timeout or
5xxdoes not prove the original task stopped. Reconcile the existing task or usage record before considering another billable action. For a task with no terminal result for an extended period, contact support with itstrace_idand task ID, never the API key.
For account-wide HTTP 429 too_many_requests recovery, follow the trial-account concurrency procedure and task-polling guidance. Changing keys or services does not bypass the trial-account slot.
Files (skills)
-
SKILL.md 2.4 KB
--- name: midjourney-image description: Explain the private Midjourney integration and diagnose existing imagine requests. Midjourney is not currently published in the AceDataCloud public catalog. license: Apache-2.0 metadata: author: acedatacloud version: "1.0" compatibility: Status and legacy-request guidance only; not a public generation workflow. The existing MCP package name is mcp-midjourney. --- # Midjourney integration status Midjourney's service and Imagine document are currently marked private. Do not submit legacy generation examples or configure `mcp-midjourney` as a public capability. An explicit API-spec lookup can still return a definition; that is not a publication or access grant. Check the current public catalog and let the user choose a published image service before changing their workflow; do not silently substitute a model or service. ## Diagnose an existing imagine request 1. Inspect the original payload locally without printing its API key. `POST /midjourney/imagine` selects a model version with `version`, not `model`. Remove the `model` key entirely, including when its value is an empty string or `null`. 2. A request containing `model` returns HTTP 400 `bad_request` before generation or asynchronous acceptance, without triggering generation or charging. For version selection the field is `"version": "8.2"`; omitting `version` keeps the default version. Correcting a payload does not authorize submitting it while the capability is private. 3. If an earlier authorized request returned a `task_id`, retain that exact ID. An accepted task is not a finished image. Only report success from the actual final `response.success: true` and usable `response.image_url` or `response.raw_image_url`; do not invent an artifact or treat a task ID as one. 4. A network disconnect, proxy timeout or `5xx` does not prove the original task stopped. Reconcile the existing task or usage record before considering another billable action. For a task with no terminal result for an extended period, contact support with its `trace_id` and task ID, never the API key. For account-wide HTTP 429 `too_many_requests` recovery, follow the [trial-account concurrency procedure](https://github.com/AceDataCloud/Skills/blob/main/skills/_shared/authentication.md#trial-account-concurrency) and [task-polling guidance](https://github.com/AceDataCloud/Skills/blob/main/skills/_shared/async-tasks.md). Changing keys or services does not bypass the trial-account slot.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.