Sources

Sources represent the data that is accessible by an application and the language model. A source can be raw text, a file (see files for supported types), a URL (website), or a set of URLs (many pages of a website).

When you create a new source, Chariot automatically extracts the text from the source (based on the source type), splits the text into chunks, embeds the chunks, and stores the resulting vectors in our vector database. This is managed for you so you don't have to maintain it in your app.

Sources are retrieved and included as relevant context for the language model during conversations with an application.

The default embedding model is OpenAI's text-embedding-ada-002. We are working on adding support for choosing your own embedding model and chunking/retrieval strategies.

The source object

Name
id
Type
string
Description
Unique identifier for the source
Name
application_id
Type
string
Description
Unique identifier for the application this source is associated with.
Name
name
Type
string
Description
Name of the source. Used when displaying where information came from in a conversation.
Name
type
Type
string
Description
Type of source. Used to determine how to process the source.
Name
content
Type
undefined
Description
Source content. This can be text, a list of URLs to crawl, or a file.
Name
embed_status
Type
string
Description
Status of the source embedding process. Once complete, your source can be queried in conversations with your application.
Name
embed_status_message
Type
string
Description
Optional message associated with the embed_status. If there is an error processing the source, this field will contain the error message.
Name
updated_at
Type
string
Description
Timestamp of when the source was last updated
Name
created_at
Type
string
Description
Timestamp of when the source was created.

GET/sources

List sources

Returns a list of all sources for your account. To get a list of all sources for an application, include the application_id query parameter.

Request

GET

/sources

curl https://api.chariotai.com/sources \
  -H "Authorization: Bearer {YOUR_API_KEY}"

Response

[
  {
        "id": "string",
        "application_id": "app_123",
        "name": "string",
        "type": "string",
        "content": null,
        "embed_status": "PENDING",
        "embed_status_message": "string",
        "updated_at": "2023-05-21T23:59:59",
        "created_at": "2023-05-19T23:59:59"
  }
]

POST/sources

Create source

Creates a new source.

Required attributes

Name
application_id
Type
string
Description
Unique identifier for the application this source should be associated with.
Name
name
Type
string
Description
Name of the source. Used when displaying where information came from in a conversation.
Name
type
Type
string
Description
Type of source. Used to determine how to process the source.
Name
content
Type
undefined
Description
Source content. This can be text, a list of URLs to crawl, or a file.

Request

POST

/sources

curl https://api.chariotai.com/sources \
  -H "Authorization: Bearer {YOUR_API_KEY}"

Response

{
  "id": "string",
  "application_id": "app_123",
  "name": "string",
  "type": "string",
  "content": null,
  "embed_status": "PENDING",
  "embed_status_message": "string",
  "updated_at": "2023-05-21T23:59:59",
  "created_at": "2023-05-19T23:59:59"
}

GET/sources/{id}

Get source

Retrieves a single source.

Request

GET

/sources/{id}

curl https://api.chariotai.com/sources/{id} \
  -H "Authorization: Bearer {YOUR_API_KEY}"

Response

{
  "id": "string",
  "application_id": "app_123",
  "name": "string",
  "type": "string",
  "content": null,
  "embed_status": "PENDING",
  "embed_status_message": "string",
  "updated_at": "2023-05-21T23:59:59",
  "created_at": "2023-05-19T23:59:59"
}

DELETE/sources/{id}

Delete source

Permanently deletes the specified source. This cannot be undone.

Request

DELETE

/sources/{id}

curl https://api.chariotai.com/sources/{id} \
  -H "Authorization: Bearer {YOUR_API_KEY}"