Training Data

Endpoints for managing training data sources used by Pylon AI.

Use Training Data to connect external knowledge sources that power Pylon’s AI.

For UI setup and behavior, see Training Data.

This API Reference section renders from our OpenAPI spec.

If you don’t see any endpoints below, they may not be published in the spec yet.

Endpoints

List all training data configurations

get

Returns all training data configurations for the organization.

Rate limit: 60 requests per minute

Authorizations

bearerAuth

AuthorizationstringRequired

Authorization: Bearer

Responses

200Success

application/json

400Error

application/json

500Error

application/json

get

/training-data

HTTPcURLJavaScriptPython

GET /training-data HTTP/1.1
Host: api.usepylon.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*

{
  "data": [
    {
      "created_at": "text",
      "documents": [
        {
          "external_id": "text",
          "file_name": "text",
          "file_size": 1,
          "file_type": "text",
          "id": "text",
          "s3_key": "text"
        }
      ],
      "exclude_from_kb_search": true,
      "id": "text",
      "name": "text",
      "scrape_status": {
        "last_scraped_at": "text",
        "pages_scraped": 1,
        "status": "text",
        "total_pages": 1
      },
      "type": "text",
      "updated_at": "text",
      "visibility": "text"
    }
  ],
  "request_id": "text"
}

Create a training data configuration

post

Creates a new training data configuration for the organization.

Rate limit: 10 requests per minute

Authorizations

bearerAuth

AuthorizationstringRequired

Authorization: Bearer

Body

application/json

training_data_namestringOptional

The name of the training data container.

visibilitystringOptional

The visibility of the training data. Defaults to "everyone".

Valid values: "everyone", "user_only", "ai_agent_only"

Responses

200Success

application/json

400Error

application/json

500Error

application/json

post

/training-data

HTTPcURLJavaScriptPython

POST /training-data HTTP/1.1
Host: api.usepylon.com
Authorization: Bearer YOUR_SECRET_TOKEN
Content-Type: application/json
Accept: */*
Content-Length: 49

{
  "training_data_name": "text",
  "visibility": "text"
}

{
  "data": {
    "created_at": "text",
    "documents": [
      {
        "external_id": "text",
        "file_name": "text",
        "file_size": 1,
        "file_type": "text",
        "id": "text",
        "s3_key": "text"
      }
    ],
    "exclude_from_kb_search": true,
    "id": "text",
    "name": "text",
    "scrape_status": {
      "last_scraped_at": "text",
      "pages_scraped": 1,
      "status": "text",
      "total_pages": 1
    },
    "type": "text",
    "updated_at": "text",
    "visibility": "text"
  },
  "request_id": "text"
}

Get a training data configuration by its ID

get

Returns a single training data configuration by ID.

Rate limit: 60 requests per minute

Authorizations

bearerAuth

AuthorizationstringRequired

Authorization: Bearer

Path parameters

idstringRequired

The ID of the training data.

Responses

200Success

application/json

400Error

application/json

404Error

application/json

get

/training-data/{id}

HTTPcURLJavaScriptPython

GET /training-data/{id} HTTP/1.1
Host: api.usepylon.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*

{
  "data": {
    "created_at": "text",
    "documents": [
      {
        "external_id": "text",
        "file_name": "text",
        "file_size": 1,
        "file_type": "text",
        "id": "text",
        "s3_key": "text"
      }
    ],
    "exclude_from_kb_search": true,
    "id": "text",
    "name": "text",
    "scrape_status": {
      "last_scraped_at": "text",
      "pages_scraped": 1,
      "status": "text",
      "total_pages": 1
    },
    "type": "text",
    "updated_at": "text",
    "visibility": "text"
  },
  "request_id": "text"
}

Upload files as training data

post

Uploads one or more files to a training data configuration.

Rate limit: 10 requests per minute

Authorizations

bearerAuth

AuthorizationstringRequired

Authorization: Bearer

Body

multipart/form-data

external_idsstring[]Optional

External IDs for idempotent operations. Must match the number of files if provided.

Used for duplicate detection and updates.

file_namesstring[]Optional

Custom names for the uploaded files. Must match the number of files if provided.

If not provided, original filenames will be used.

filesstring[]Required

The files to upload (minimum 1, maximum 10).

Supported types: PDF, plain text, markdown, CSV, JSON, and images (JPEG, PNG, GIF, WebP).

Maximum size: 50MB per file, 100MB total.

training_data_idstringOptional

The ID of an existing training data container to upload files to.

Required if not creating new training data.

training_data_namestringOptional

The name of a new training data container.

Required when creating new training data (if training_data_id is not provided).

visibilitystringOptional

Visibility of the training data. Defaults to "everyone".

Valid values: "everyone", "user_only", "ai_agent_only"

Responses

200Success

application/json

400Error

application/json

500Error

application/json

post

/training-data/upload

HTTPcURLJavaScriptPython

POST /training-data/upload HTTP/1.1
Host: api.usepylon.com
Authorization: Bearer YOUR_SECRET_TOKEN
Content-Type: multipart/form-data
Accept: */*
Content-Length: 138

{
  "external_ids": [
    "text"
  ],
  "file_names": [
    "text"
  ],
  "files": [
    "text"
  ],
  "training_data_id": "text",
  "training_data_name": "text",
  "visibility": "text"
}

{
  "data": {
    "created_at": "text",
    "documents": [
      {
        "external_id": "text",
        "file_name": "text",
        "file_size": 1,
        "file_type": "text",
        "id": "text",
        "s3_key": "text"
      }
    ],
    "exclude_from_kb_search": true,
    "id": "text",
    "name": "text",
    "scrape_status": {
      "last_scraped_at": "text",
      "pages_scraped": 1,
      "status": "text",
      "total_pages": 1
    },
    "type": "text",
    "updated_at": "text",
    "visibility": "text"
  },
  "request_id": "text"
}

Upload text content as training data

post

Uploads plain text content as a training data document.

Authorizations

bearerAuth

AuthorizationstringRequired

Authorization: Bearer

Body

application/json

contentstringOptional

The text content to upload.

external_idstringOptional

External ID for idempotent operations (optional).

file_namestringOptional

The name for the uploaded file.

training_data_idstringOptional

The ID of an existing training data container to upload to (optional).

training_data_namestringOptional

The name of a new training data container (required if training_data_id not provided).

visibilitystringOptional

Visibility of the training data. Defaults to "everyone".

Responses

200Success

application/json

400Error

application/json

500Error

application/json

post

/training-data/upload-content

HTTPcURLJavaScriptPython

POST /training-data/upload-content HTTP/1.1
Host: api.usepylon.com
Authorization: Bearer YOUR_SECRET_TOKEN
Content-Type: application/json
Accept: */*
Content-Length: 132

{
  "content": "text",
  "external_id": "text",
  "file_name": "text",
  "training_data_id": "text",
  "training_data_name": "text",
  "visibility": "text"
}

{
  "data": {
    "created_at": "text",
    "documents": [
      {
        "external_id": "text",
        "file_name": "text",
        "file_size": 1,
        "file_type": "text",
        "id": "text",
        "s3_key": "text"
      }
    ],
    "exclude_from_kb_search": true,
    "id": "text",
    "name": "text",
    "scrape_status": {
      "last_scraped_at": "text",
      "pages_scraped": 1,
      "status": "text",
      "total_pages": 1
    },
    "type": "text",
    "updated_at": "text",
    "visibility": "text"
  },
  "request_id": "text"
}

Delete training data documents

delete

Removes documents from a training data configuration by document ID or external ID.

Rate limit: 10 requests per minute

Authorizations

bearerAuth

AuthorizationstringRequired

Authorization: Bearer

Path parameters

idstringRequired

The ID of the training data.

Query parameters

document_idsstring[]Optional

Document IDs to delete.

external_idsstring[]Optional

External IDs to delete.

Responses

200Success

application/json

400Error

application/json

404Error

application/json

500Error

application/json

delete

/training-data/{id}/documents

HTTPcURLJavaScriptPython

DELETE /training-data/{id}/documents HTTP/1.1
Host: api.usepylon.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*

{
  "deleted_count": 1,
  "request_id": "text"
}