EvoLink Credits (余额查询)

Every EvoLink generation node in this pack charges credits. A 4K Kling clip can eat a real chunk of your balance, and nothing stings quite like watching a job fail at minute three with a 402 because you ran out mid-queue. That's the entire reason EvoLink_Credits exists: it's a free read-only call to check what's left before you commit to an expensive run.

The node is as simple as it gets - it has no inputs you're required to fill and it doesn't generate anything. It just asks the EvoLink API for your remaining balance and hands it back to you.

What it does

Behind the scenes it calls the GET /v1/credits endpoint, which is free and never consumes credits - the check itself won't cost you anything, even when your balance is nearly zero. The node's one deliberate quirk: it refuses to use ComfyUI's cache (IS_CHANGED always returns a change), so it re-queries every single time you run it. Most generator nodes skip work when nothing changed; this one is designed to never do that, because your balance changes with every job you run.

Two outputs come out of it:

• remaining_credits (FLOAT) - your account balance as a number. Wire this into anything that wants a number, or just read it.
• info (STRING) - a human-readable summary: account balance, and if your current key has its own spending limit set on the dashboard, that key's remaining quota too, plus a link to top up.

Using it

There's one optional input, api_key, and it behaves exactly like every other node in the pack: fill it the first time and it's saved to a local config file (evolink_config.json in the pack folder, chmod 600) so you can leave it blank forever after. Or skip it entirely and set the EVOLINK_API_KEY environment variable instead.

For the workflow-level move, stick it at the front of a queue with an API wrapper or a Show Text node on info, and run it once before you fire off a batch of video jobs. It's also genuinely useful for learning what a model costs: run a node, then glance at its status_info output - it lists credits used per task - and you'll build an instinct for which tier is the budget pick.

Installing and troubleshooting

Same story as the rest of the pack - no GPU, no model files, the only Python dependency is requests:

cd ComfyUI/custom_nodes
git clone https://github.com/deeplearning-goethe/comfyui-evolink.git

then fully restart ComfyUI (close the console window, not just the browser tab). Or use ComfyUI Manager and search EvoLink.

If you get a 401, the key is invalid or was deleted. 402 means empty balance, which is exactly the situation this node exists to catch - go top up. One thing to know going in: this is a young, third-party pack, and the "official" in the name means official from EvoLink, not a Comfy Org partner node. There's not much community history behind it yet, so keep an eye on the key: it's stored locally on your machine and should never end up in a shared workflow JSON - clear the api_key field before you export.