suggest_tags

Suggest NovelAI prompt tags that complete or refine a partial prompt. The suggestions are ranked by NovelAI’s training-set tagger and returned with their training occurrence count.

Parameters

Name

Type

Default

Notes

prompt

str

required

Partial prompt text to complete.

model

str

"nai-diffusion-4-5-full"

Tagger vocabulary (matches a Model enum value).

language

str

"en"

ISO 639-1 response language (e.g. en, ja, zh).

Returns

list[TagSuggestion] — each suggestion carries the fields NovelAI returns:

Key

Type

Notes

tag

str

The tag string itself (use this in your prompt).

count

int

Training-set occurrence count (higher = commoner).

confidence

float

Tagger confidence for this completion (0–1).

The wire shape is verified live (see API surface alignment):

Example

result = await ctx.session.call_tool("suggest_tags", {
    "prompt": "1girl, fox ears,",
    "model": "nai-diffusion-4-5-full",
    "language": "en",
})
# result: [
#   {"tag": "fox girl", "count": 10000, "confidence": 0.745},
#   {"tag": "fox tail", "count": 4321, "confidence": 0.612},
#   ...
# ]

Tips

Tip

Iterative prompt building: ask an agent to call suggest_tags after each prompt edit, then accept the top suggestion that fits the scene. This is a fast way to discover tags you wouldn’t have thought of.

Tag count vs quality

count measures how often the tag appears in NovelAI’s training data — not how often it’s useful for your prompt. A rare tag (low count) can still be the perfect descriptor for your scene.

See also