Skip to main content
POST
/
v1
/
detokenize
Detokenization
curl --request POST \
  --url http://localhost:8000/v1/detokenize \
  --header 'Content-Type: application/json' \
  --data '
{
  "tokens": [
    128000,
    3923,
    374,
    1803,
    1413,
    15592,
    30
  ]
}
'
{
  "text": "What is generative AI?"
}
By giving a list of tokens, generate a detokenized output text string.

Body

application/json
​
tokens
integer[]
required

A token sequence to detokenize.

Example:
[128000, 3923, 374, 1803, 1413, 15592, 30]
​
model
string | null

Routes the request to a specific adapter.

Example:

"(adapter-route)"

Response

Successfully detokenized the tokens.

​
text
string
required

Detokenized text output.

Last modified on June 14, 2026