const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    description: 'men\'s cotton t-shirt',
    predicted_hs_code: '6109.10.0012',
    correct_hs_code: '6109.10.0012',
    feedback_type: 'positive',
    mode: 'pro'
  })
};

fetch('https://api.traddal.com/v1/classify/feedback', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "success": true,
  "feedback_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "feedback_type": "negative",
  "created_at": "2026-04-02T10:00:00.000Z"
}

{
  "success": false,
  "request_id": "c6f398bb-1dad-40e8-a991-a22a180ad842",
  "error": "invalid_request",
  "message": "Invalid request parameters"
}

{
  "success": false,
  "request_id": "c6f398bb-1dad-40e8-a991-a22a180ad842",
  "error": "invalid_request",
  "message": "Invalid request parameters"
}

Classification

Submit Classification Feedback

Submit feedback on a classification result. Use this to confirm a classification was correct (positive) or to provide the correct HS code when it was wrong (negative).

This data is used to improve classification accuracy over time.

Feedback types (feedback_type):

positive — The predicted HS code was correct
negative — The predicted HS code was incorrect (provide the correct code)

Reason types (reason_type): (required for negative feedback)

wrong_code — Wrong HS code
wrong_chapter — Wrong chapter entirely
close_but_wrong — Close but wrong subheading
material_mismatch — Wrong material classification
missing_context — Needs more context to classify
other — Other (provide details in reason field)

POST

/

v1

/

classify

/

feedback

const options = {
  method: 'POST',
  headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
  body: JSON.stringify({
    description: 'men\'s cotton t-shirt',
    predicted_hs_code: '6109.10.0012',
    correct_hs_code: '6109.10.0012',
    feedback_type: 'positive',
    mode: 'pro'
  })
};

fetch('https://api.traddal.com/v1/classify/feedback', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "success": true,
  "feedback_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "feedback_type": "negative",
  "created_at": "2026-04-02T10:00:00.000Z"
}

{
  "success": false,
  "request_id": "c6f398bb-1dad-40e8-a991-a22a180ad842",
  "error": "invalid_request",
  "message": "Invalid request parameters"
}

{
  "success": false,
  "request_id": "c6f398bb-1dad-40e8-a991-a22a180ad842",
  "error": "invalid_request",
  "message": "Invalid request parameters"
}

Authorizations

Authorization

string

header

required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json

description

string

required

The product description that was classified

Example:

"men's cotton t-shirt"

predicted_hs_code

string

required

The HS code returned by the classification

Example:

"6109.10.0012"

correct_hs_code

string

required

The correct HS code. Same as predicted_hs_code for positive feedback.

Example:

"6109.90.1007"

feedback_type

enum<string>

default:negative

Whether the classification was correct (positive) or incorrect (negative)

Available options:

positive,

negative

reason_type

enum<string>

Category of why the classification was incorrect:

wrong_code — Wrong HS code
wrong_chapter — Wrong chapter entirely
close_but_wrong — Close but wrong subheading
material_mismatch — Wrong material classification
missing_context — Needs more context to classify
other — Other (provide details in reason field)

Available options:

wrong_code,

wrong_chapter,

close_but_wrong,

material_mismatch,

missing_context,

other

reason

string

Freetext explanation (used when reason_type is "other")

Example:

"Product is actually a blend, not 100% cotton"

mode

enum<string>

The classification mode that produced the original result

Available options:

mini,

pro,

max

Response

Feedback submitted successfully

success

boolean

Example:

true

feedback_id

string<uuid>

Unique ID of the feedback record

Example:

"a1b2c3d4-e5f6-7890-abcd-ef1234567890"

feedback_type

enum<string>

Available options:

positive,

negative

Example:

"negative"

created_at

string<date-time>

Example:

"2026-04-02T10:00:00.000Z"