Midjourney by Ace Data Cloud

Midjourney for ComfyUI by Ace Data Cloud

Access limitation (2026-10-19): the live Ace service catalog marks this Midjourney service private=true. This guide maintains existing authorized workflows; it does not promise new-account access or announce a public service launch. Registry package Active does not establish service entitlement. Confirm your existing application permission before any paid action. No new public promotion/package release is requested by this maintenance update.

Maintain an existing authorized Midjourney workflow using action-specific fields and the same accepted task ID. This independent package calls a hosted API; no local model or GPU is needed. Check your access and current service pricing before any paid action.

简体中文 · Ace Data Cloud · Issues

Install

Open the official Midjourney by Ace Data Cloud Registry page, or search Midjourney in Manager → Custom Nodes Manager. Verify the exact Ace publisher/repository below; use Ace Data Cloud as a fallback if the service spelling does not return it. Select Midjourney by Ace Data Cloud, confirm publisher acedatacloud and source repository AceDataCloud/MidjourneyComfyUI, then click Install. In Select Version, explicitly select the approved Registry stable release 0.1.0. The normal dialog may initially highlight nightly; do not accept it for this stable-release tutorial. Keep dependency installation enabled. After Restart Required, restart ComfyUI and refresh the browser. Registry ID: midjourney.

These are real 2026-10-16 Linux ComfyUI0.41.0 brand-search/installed-state screenshots, not a new service-term search or a paid request. See sources and limits.

This Registry stable installation was verified on 2026-10-16 with original ComfyUI 0.41.0 on Linux CPU. That test did not submit a generation or chat request. Newer host versions and each paid action require their own checks; model availability and price remain subject to the current service rules.

For local source testing, clone this repository into ComfyUI/custom_nodes/MidjourneyComfyUI and run python -m pip install -r requirements.txt in ComfyUI's environment, then restart. Native media loaders require ComfyUI 0.39.0 or newer. This package uses ComfyUI's existing image/audio/video libraries for decoding.

Key and tool inputs

Copy or create your API Token

Sign in/register and initialize Applications → General application if needed. Copy an existing permitted API Token, or use Manage Keys → Create for a dedicated ComfyUI key. Check service access, expiry/Allowed APIs, Credits balance and current price first.

These masked 2026-10-08 Ace Console images are reused from SunoDify. The sample name Dify tutorial should be ComfyUI here; no credential was created for this guide. Configure only the Token string in the actual process that launches ComfyUI (ACEDATACLOUD_API_KEY), then restart. There is no Key widget in these nodes; never paste it into a prompt, JSON parameter, workflow or screenshot. An export in another terminal does not configure an already-running Desktop process.

1. Sign in to Ace Data Cloud → Applications. If your account has no General application, choose Initialize General Application.
2. Open General application and check access to the service used by this package. A service-specific application key is limited to that service. Copy an existing API key, or choose Manage Keys → Create for a separate key named ComfyUI.
3. Copy only the token string, without Bearer or quotation marks. A platform management token is not the API key these nodes use. If you enable Allowed APIs, include every submission and task-query path used by your workflow. Check current pricing and available Credits before a paid action.

For a terminal launch, set the Token privately and start ComfyUI from that same terminal:

export ACEDATACLOUD_API_KEY='paste-your-private-token-here'
python main.py --cpu

Set ACEDATACLOUD_API_KEY (or ACEDATACLOUD_API_TOKEN) in the ComfyUI process environment and restart. Keep the key out of workflow JSON, screenshots, issues, and the repository.

Each Dify tool has its own node. The model and action selectors expose every fixed choice in the source YAML. Other fields appear as named text or confirmation widgets; leave one empty to omit it. parameters_json accepts the same API fields as an advanced JSON object. A nonempty widget overrides the matching JSON field; (from JSON/default) keeps a model/action supplied by JSON or the API default. Arrays and objects can be entered as JSON strings in their named widgets. Task lookup nodes also offer task_id and wait_seconds; batch lookup nodes offer ids_json. Contract validation happens before a request. Paid submissions are never retried after a connection failure; preserve an accepted task ID and inspect request history before paying again.

| Dify tool | Required fields in JSON | Model choices | Action choices | | --- | --- | --- | --- | | midjourney_describe | image_url | — | — | | midjourney_edits | — | — | — | | midjourney_get_seed | image_id | — | — | | midjourney_imagine | — | — | — | | midjourney_shorten | prompt | — | — | | midjourney_task_retrieve | — | — | — | | midjourney_tasks_retrieve_batch | — | — | — | | midjourney_translate | content | — | — | | midjourney_videos | — | — | generate, extend |

Import 01-configure-and-run.json with File → Open, fill its placeholder task ID or JSON parameters, then run. The example is a configuration template; it does not submit a paid request by itself. Query the same task ID until status is succeeded. The node returns status, task ID, all media URLs as JSON, first media URL, normalized result JSON, and trace ID. Connect a completed Ace CDN URL to the provided IMAGE, VIDEO loader where applicable.

Check current service pricing before a paid run. Read-only task queries and media downloads should not generate a new paid task. Destructive Dify actions still require an explicit confirm: true in JSON. See privacy notes.

Same-task refresh and cancellation limits

Disable auto-queue/auto-run before importing paid submission workflows and submit once. This package version does not declare a task-cache freshness hook or an interruptible polling callback. If an unchanged retrieve graph returns an earlier pending result, change wait_seconds (for example240 to239) and run only retrieval for the same task ID to invalidate the cache. Do not rerun generation. Stop may wait for the local poll/HTTP timeout; it does not cancel or refund an accepted remote task. A single retrieve with wait_seconds=0 is useful for explicit status checks. The separate Top5 runtime fixes are not claimed for this package.

Troubleshooting

| Symptom | Action | | --- | --- | | Missing Key /401 | Configure the actual ComfyUI launch-process API Token and restart; a platform management Token is not an API Key. | | 403 | Read the returned reason and check service entitlement, allowed paths, expiry or content restriction; do not blindly resubmit. | | 429 | Wait for rate limits; keep any accepted task ID and avoid duplicate paid submissions. | | Invalid JSON/model/input | Correct the named field according to the tool table; {}/example URLs are placeholders, not a complete paid request. | | Submit timeout/5xx | Inspect request history/trace ID before another paid submission. | | Unexpected result or media failure | Follow this service's synchronous/async section, confirm a completed real media URL where supported, and inspect Console. No result is fabricated from pending status. |

Development

python -m unittest discover -s tests -v
python -m py_compile __init__.py client.py api_contracts.py nodes.py

pyproject.toml declares short Registry ID midjourney, publisher acedatacloud, and a 400×400 proportional resize of the canonical Ace logo.