List Available Models
The little node that tells you what API models you can call
- model_list
This is a small utility node, and it does one honest thing: it asks the AI providers you've configured which models your account can actually use right now, and hands you back that list as text. When you're staring at a hard-coded dropdown on the Gemini or OpenAI node wondering whether gemini-2.5-pro is even available on your key, this is how you check without leaving ComfyUI.
Why you'd want it
API model names churn. Providers rename models, retire old ones, gate the newest behind tiers, and roll out previews that only some accounts see. The dropdowns on the pack's main API nodes are baked in when the pack was published, so they drift out of date - a name in the menu might 404 against your key, and a shiny new model won't be in the menu at all. ListAvailableModels sidesteps that by querying the provider live with your credentials, so what comes back is the ground truth for your account, not a stale hard-coded list. It's a debugging and discovery tool: run it when a model call fails with "model not found," or when you want to know what your key unlocks.
How it works
It calls the "list models" endpoint on the providers you switch on and concatenates the results into one string. That's the whole mechanism - no generation, no image, no creativity dials. It reads the same API keys from the pack's config that the other nodes use, so if a provider returns nothing, that usually means the key is missing or wrong for that provider.
The inputs and outputs that matter
There are only two inputs, both simple toggles:
display_gemini(default on) - include the models your Gemini/Google key can see.display_openai(default on) - include the models your OpenAI key can see.
Flip either off to narrow the list to the provider you care about. The single output is model_list, a string - the names, ready to read. Wire it into a Show Text-style preview node (or the pack's Save Text node) so you can actually see it; a string output with nowhere to go doesn't display itself.
Installing it
It comes with the pack. ComfyUI Manager: search ComfyUI-OllamaGemini, install, restart. Or:
cd ComfyUI/custom_nodes
git clone https://github.com/al-swaiti/ComfyUI-OllamaGemini
pip install -r requirements.txt
then restart. No model files, no extra dependencies beyond what the pack already installs. It does, however, need your API keys configured - this node is only as useful as the credentials it can query with.
Where people get burned
Two small things. First, an empty or partial list almost always means a key problem, not a node bug: if display_gemini is on but nothing Google-shaped comes back, your GEMINI_API_KEY isn't set or isn't valid. Fix the key, not the node. Second, the output is just a string - beginners wire it into nothing and then wonder why the node "does nothing." Attach a text-preview node to model_list so the answer is actually on screen. Beyond that, it's about the most harmless node in the pack: it reads, it never generates, and it costs nothing but a lightweight API call.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| display_gemini | BOOLEAN | true | — |
| display_openai | BOOLEAN | true | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| model_list | STRING | — |