Griptape Agent Config: OpenAI [DEPRECATED]

This is the original OpenAI structure config - the node that powered most early Griptape-in-ComfyUI workflows. It's marked deprecated now because the newer "Griptape Agent Config: OpenAI Drivers" node does strictly more in a cleaner way, but the old one is still out there in thousands of saved workflows, still loads, and still runs fine. If you've just opened a workflow you found online and this is sitting in it, here's what it is and whether you should care.

What it is

A single config node that points a Griptape agent's prompt driver at OpenAI. One input is required: prompt_model - a free-text model name (defaults aren't pre-filled, so you type something like gpt-4o-mini). Optional inputs are the old-style set: temperature, seed, max_attempts_on_fail, env, use_native_tools (default on), max_tokens, and api_key_env_var (default OPENAI_API_KEY). Output is a single CONFIG.

Mechanically it builds an OpenAI chat driver from the model name and key you supply, wraps it in a DriversConfig, and hands it to the agent. It does not set up image generation, embeddings, or audio - those came later, in the Drivers version.

Why it's deprecated

The replacement, OpenAI Drivers, bundles everything: prompt driver with a proper 60+ model dropdown, image generation driver (dall-e-2/dall-e-3/gpt-image-1), embedding driver, text-to-speech driver, and audio transcription driver, all in one config. This old node is prompt-only, which meant memory, image, and audio features all had to be assembled from separate nodes. It's kept so legacy workflows keep loading.

Installing it

It ships in the one pack:

cd ComfyUI/custom_nodes
git clone https://github.com/griptape-ai/ComfyUI-Griptape

or ComfyUI Manager → search "Griptape". Restart, set your OPENAI_API_KEY in Settings → Griptape (or as an env var), and old workflows that reference this node will run.

Should you use it?

Only if you're preserving an existing workflow. The migration to OpenAI Drivers is as easy as these swaps get: same CONFIG socket, and the new node defaults to gpt-4o-mini with dropdowns instead of free text - you re-pick your model and gain image/embedding/audio support you didn't have. New work should use the Drivers config, period.

Common issues

• prompt_model not set - it's required and free-text; a typo produces an API error at runtime.
• No memory, no image gen, no audio - expected, it's the old prompt-only design.
• [WARNING]: OPENAI_API_KEY is not set - the standard pack key warning; check Settings → Griptape and env var spelling. This is the single most common failure across the whole pack.