Kling API Health Check
Why Isn't My Kling Workflow Running? Check Here First
- auth
- is_healthy
- status
Kling API Health Check answers the single most annoying question in the ComfyUI-Kling-Direct pack: is it me, or is it Kling? Every generation node in this pack fails in the same ambiguous way when something's wrong up the chain - bad keys, dead endpoint, KYC not finished - so a one-click triage node is genuinely useful, not just decoration. It's the first thing I drop into any new Kling workflow when something stops working.
It's a v2.1 addition, which tells you the author hit the same debugging wall the rest of us do. The pack's README positions it in exactly those terms: "Use it to debug 'why isn't my workflow running.'"
How it works
The node takes one input - auth (the KLING_AUTH object from the Kling AI Authentication node) - and makes a single cheap API call to Kling's effect-templates endpoint. That endpoint is effectively free, so it's a safe smoke test. It returns:
is_healthy(BOOLEAN) -trueif the request succeeded,falseif anything threw.status(STRING) - a human-readable message. Success says something like "auth + connectivity verified"; failure carries the API error or the connection exception type.
The source is refreshingly simple: try the call, return the boolean, catch KlingAPIError and generic exceptions into the status string. No polling, no state, no downloads. Because it has the AlwaysExecuteMixin, it runs even when ComfyUI thinks nothing changed downstream - which is what you want from a test node. Wire it up on its own and hit "Run" whenever you're suspicious.
How to actually use it
The node is designed to be run standalone, but if you want it in a workflow permanently you can feed it the same auth your generation nodes use and treat is_healthy as a gate. Realistically it's a debugging tool: run it, read the status, fix the thing, move on.
The status message separates the failure modes for you, which is the whole point:
- "Kling API ERROR: ..." - the request reached Kling and Kling rejected it. That's a credentials problem: wrong keys, expired secret, or an account whose KYC activation never completed. Re-check the keys in your Auth node.
- "Kling API connection failed: ..." - you never got a response at all. That's DNS, network, a proxy, or a wrong region endpoint. If you're hitting the default
api-singapore.klingai.comand you're not in Singapore, consider whether the Kling Region Selector should be pointing somewhere else.
Common gotchas
- Healthy here doesn't mean everything downstream works. This only checks auth + connectivity to one endpoint. Your text-to-video call can still fail on its own terms - bad prompt formatting, quota exhausted, an invalid parameter. Health check is a first filter, not a guarantee.
- You still need a valid
auth. There's no env-var shortcut on this node; it only takes the auth object. If you can't produce one, that's your problem, and honestly, the health check will just tell you. - Don't forget the KYC step. If you just created keys and the check says the API rejected you, the single most common cause in the pack is an account that hasn't finished KYC activation at
https://app.klingai.com/global/dev. The health check doesn't tell you that directly - it just reports the error.
Install
It ships in ComfyUI-Kling-Direct, so it comes with the pack: ComfyUI Manager → Install Custom Nodes → search "Kling Direct" → install, restart. Or:
cd ComfyUI/custom_nodes
git clone https://github.com/IxMxAMAR/ComfyUI-Kling-Direct
No extra dependencies, no model files, no GPU needed. If the pack is already installed, you already have this node - you just haven't had a reason to use it yet. Give it one the next time a Kling node fails in a way you can't immediately explain.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| auth | KLING_AUTH | — |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| is_healthy | BOOLEAN | — |
| status | STRING | — |