⚠️ [Deprecated] Generate Image (from Text)

If you loaded an old DeepGen workflow and got a node named ⚠️ [Deprecated] Generate Image (from Text), don't panic and don't delete it. This is the same node as DeepGen_T2I0 - literally the same Python class underneath - wearing its pre-rename name and parked in the DeepGen/Deprecated category so you can see at a glance that it's the legacy spelling.

Why it exists

Back around pack version 1.2.1 the author reworked the naming scheme. All the node keys got padded to a uniform length and the current-generation classes picked up 0 (or X) suffixes - DeepGen_T2I0, DeepGen_I2I0, DeepGen_T2T0, and so on. Rather than break every workflow that referenced the old class names, the pack registers both: the new names point at the current classes, and the old names are registered as thin subclasses that just override the category and set ComfyUI's DEPRECATED flag. Your old graph keeps executing without edits, and you get a warning triangle instead of a missing-node error.

That's the whole difference between DeepGen_T2I and DeepGen_T2I0: category (DeepGen/Deprecated vs DeepGen/Generators) and display name. The inputs, outputs, API calls, and behavior are identical.

What it does

Text-to-image, cloud-rendered via the DeepGen gateway - see the DeepGen_T2I0 article for the full breakdown. You pick a model (Nano Banana Pro, Nano Banana 2, GPT Image 1.5, Grok Imagine, FLUX.1 [schnell], Seedream, Wan 2.6, Hunyuan Image 3.0), write a prompt, choose nb_results, minimum_resolution, aspect_ratio, output_format, and it returns an IMAGE batch, an output_prefix_and_model string, and a total_credits_used float. Same key requirement (Settings → "DeepGen API Key"), same credit metering, same cloud round-trip.

Install

It's the same single pack either way:

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, then add your DeepGen API key from deepgen.app in Settings - saved to ComfyUI/user/deepgen/config.json.

What you should actually do

For new workflows, use DeepGen_T2I0. The deprecated node still works and is kept alive precisely so legacy graphs don't rot, but every new name is pointing you somewhere. When you open an old graph and see the warning flag, the cleanest fix is to swap the node for DeepGen_T2I0 and reconnect - the sockets are identical, so it's a two-second replacement. If you've got a stable old workflow you never touch, leaving it as-is is perfectly fine too. "Deprecated" here means "renamed," not "broken."