Ingest Knowledge (alpha)

Ingest knowledge into the copilot.

POST /backend/knowledge/:bot_id

Example Request

curl --location 'http://localhost:8888/backend/knowldgebase/:bot_id' \
  --header 'Authorization: Bearer TOKEN' \
  --header 'Content-Type: application/json' \
  --data '{
      "sources": ["https://en.wikipedia.org/wiki/Huffman_coding", "https://example.com/file.pdf"]
  }'

Response

[
    {
       "message": "Knowledge has been queued for ingestion"
    }
]

Important: This endpoint is in alpha and may change in the future. please report any issues or feedback to the issue tracker.

Important: this endpoint only supports URLs, if you have a local file that you want to ingest, you can upload it to a cloud storage service and provide the URL to the file.

Ingest Knowledge

Knowledge is any information that the copilot can use to answer questions. as for now, the copilot can ingest knowledge in the following formats:

• Website: when you provide a URL, the copilot will scrape the content of the website recursively and ingest the content into its knowledge base, you can also provide a list of URLs to scrape.
• Files: you can provide a list of files to ingest, the copilot will ingest the content of the files into its knowledge base, currently, the copilot supports the following file types:
  • PDF
  • Markdown

Example Request

curl --location 'http://localhost:8888/backend/knowldgebase/:bot_id' \
  --header 'Authorization: Bearer TOKEN' \
  --header 'Content-Type: application/json' \
  --data '{
      "sources": ["https://en.wikipedia.org/wiki/Huffman_coding", "https://example.com/file.pdf"]
  }'

Response

[
    {
       "message": "Knowledge has been queued for ingestion"
    }
]