Trello Browser (Advanced)

Here's a workflow problem you didn't know you had: your prompt library lives in a Trello board, each card is a prompt, and you want to click through them in ComfyUI, see the attached reference image, and generate. Trello Browser (Advanced) is the answer - a node with an actual embedded web UI that lists the cards on a board, shows you which list they're on and their image attachments, and lets you pick one to feed the graph. It's the "manual review" node from the pack's Trello suite: use it when you want a human in the loop, deciding prompt-by-prompt what gets generated.

The name is a small lie, in the good way. It doesn't call Trello's API and needs no API key. Instead it reads a Trello board export - a JSON file you save from Trello - and the node's JavaScript UI browses that file. That's the entire integration, and it means it works offline, forever, with zero credentials. The author ships the pack with a sample reference JSON (reference/jggNQEHi - prompts.json) that the json_path defaults to, so the node even has something to show before you point it at your own board.

How it works

Point json_path at your Trello JSON export. The built-in UI (a JS extension in the pack) calls a ComfyUI API route (/another_utils/trello_prompts) that parses the file: it skips archived cards, maps each card to its list, and pulls the first .png/.jpg/.jpeg/.webp attachment URL as the preview. Click a card and its ID lands in the selected_id field; on execution the node looks up the card and returns its data. There's even a refresh button, since the UI caches the parsed file.

The inputs that matter

• json_path - path to a Trello board JSON export.
• list_filter - dropdown of the board's lists, plus "All". Pre-populated when the node loads.
• selected_id - set by clicking a card in the browser UI. You don't usually type this yourself.

Three outputs: prompt (the card's name), category (which list the card lives in), and image_url (the attached image's URL, if any). Wire prompt into a CLIP Text Encode, and use image_url however you like - it's a string, so it feeds display nodes or a loader that accepts URLs, not raw pixels.

Getting your JSON export

Open your board in Trello → Menu → More → Print and Export → Export JSON. Save it somewhere and point json_path at it. That's the whole setup.

Install

Manager → search AnotherUtils, or:

cd ComfyUI/custom_nodes
git clone https://github.com/marcoc2/ComfyUI-AnotherUtils.git

Restart. No API keys, no requirements, no models.

Common issues

If the list dropdown says "No lists found (check path)", the JSON didn't parse - wrong path, malformed export, or the file can't be read (check your console for Error parsing Trello JSON). If the browser shows nothing after a path change, hit the refresh button; it loads on demand, not automatically. And remember the node reads card names as prompts - if your prompts live in card descriptions instead, you'll want the JSON edited or the Prompt Loader's sibling behavior, since this node surfaces the name.