⚠️ [Deprecated] Invoke LLM

If an old DeepGen workflow handed you a node labelled "⚠️ [Deprecated] Invoke LLM", that's DeepGen_T2T - and it's not broken, just renamed. It's the legacy spelling of the pack's cloud LLM node, which now ships as DeepGen_T2T0. Same class underneath, same inputs, same outputs, same API call. The only differences are the display name, the category (DeepGen/Deprecated), and the warning triangle, all courtesy of the pack's 1.2.1-era rename where current nodes picked up 0/X suffixes and old keys were kept as compatibility aliases.

What it does

Text in, text out, from a frontier cloud model: Claude Opus 4.6 / Sonnet 4.6, GPT-5.4, Gemini 3 Pro / Flash, Grok 4.20, Gemini 3.1 Flash Lite. Its classic job in a ComfyUI graph is prompt enhancement - take a rough idea and rewrite it into a detailed prompt for a generator node - or any general text-processing task you want wired into the graph. Outputs are output (STRING), model (STRING), and total_credits_used (FLOAT). The full breakdown is in the DeepGen_T2T0 article.

Install

Same single pack:

cd ComfyUI/custom_nodes
git clone https://github.com/deepiksdev/ComfyUI-DeepGen-API.git
cd ComfyUI-DeepGen-API
pip install -r requirements.txt

Restart ComfyUI, and drop your DeepGen API key (deepgen.app) into Settings → "DeepGen API Key" - stored at ComfyUI/user/deepgen/config.json.

What to do with it

New graphs should use DeepGen_T2T0; there's no reason to keep reaching for the deprecated spelling. Old graphs that already reference DeepGen_T2T keep executing with no edits, because that's exactly what the alias is for. If you're cleaning up a legacy workflow, swap in DeepGen_T2T0 and reconnect - the sockets match. And if a stable old graph works, leave it. "Deprecated" here is the pack's polite way of saying "renamed, but we kept your old files working," not "this will be deleted."