to DICT

Some nodes hand you a Python dict and expect one back, and some hand you the same data in a shape that's not a dict - a list of key-value pairs, say - and the connection won't happen. to DICT is the converter for that gap: it takes a compatible input and produces a proper DICT type. It's part of Basic data handling, StableLlama's zero-dependency pack, and it's honestly the most niche of the pack's cast nodes - you'll use to INT and to FLOAT all the time, but you'll reach for this one only when a workflow starts moving structured data around.

How it works

The conversion is Python's dict(input), and that's where the constraints come from. A dict can only be built from:

• a mapping - something that's already dict-like, or
• an iterable of key-value pairs - e.g. a list of two-element pairs like [["name", "Anna"], ["steps", 30]].

Anything else fails, and the node fails loudly: it catches the invalid case and raises a ValueError with a message telling you the input must be a mapping or a list of key-value pairs. That's good behavior, honestly - a silent garbage dict would be worse. What you can't feed it: a plain list of scalars, a string, a number. Those have no key-value structure, and no amount of wishing will make one.

The input and the output

• input (ANY) - the mapping or list of pairs to convert.
• DICT (output) - the resulting dictionary.

That's the whole node. Downstream, whatever expects a DICT socket will accept it; the pack's dict nodes (get, set, keys, values, merge...) are the natural consumers.

Installing it

Install Basic data handling from ComfyUI Manager (search "Basic data handling"), or:

cd ComfyUI/custom_nodes
git clone https://github.com/StableLlama/ComfyUI-basic_data_handling

Restart ComfyUI. No dependencies, no model files.

Troubleshooting

• "It throws 'Cannot convert... to a DICT'." Expected when the input isn't a mapping or a list of pairs. The most common real-world case: someone built a list of plain values and expected a dict to materialize. A dict needs keys and values - if your data has no keys, it genuinely isn't dict-shaped, and you should restructure upstream rather than fight the cast.
• "I fed it a list of pairs and it still failed." Check the pairs are actually two elements each - a single-element pair or a ragged list breaks dict(). Also, duplicate keys don't error, they just silently keep the last one, which is worth knowing when you're debugging a missing value.
• "I wanted a JSON string, not a dict." Different conversion. This node produces the DICT object type, not a string. Look at the pack's string or regex nodes (or a JSON utility pack) if you need serialized text.