Add Text-type Documents

Batch upload text-typedocuments, which will be processed sequentially through chunking/slicing, embedding/vectorization, and storage to obtain new document IDs.

Note:
You can specify the embedding model with embedding_model_version_id; if omitted, the system default embedding model is used.
Only upload results are returned, not the final embedding results. You can obtain the final results through the "Query Document Status" API.

Request Method

POST

Request URL

https://api-${endpoint}.gptbots.ai/v1/bot/doc/text/add

Request Authentication

See Overview for authentication details.

Request

Request Example

curl -X POST 'https://api-${endpoint}.gptbots.ai/v1/bot/doc/text/add' \ -H 'Authorization: Bearer ${API Key}' \ -H 'Content-Type: application/json' \ -d '{ "knowledge_base_id": "67457fea6f658672d6482542", "embedding_model_version_id": "6620f06262390a0be1411c5d", "chunk_token": 700, "splitter": "\n", "files": [ { "file_url": "https://www.gptbots.ai/docs/article_1.pdf", "file_base64": "SGVsbG8sIEJhc2U2NCBFbmNvZGluZyE=", "source_url": "https://www.gptbots.ai/docs/article_1.pdf", "file_name": "article_1.pdf" } ] }'

                      
                      curl -X POST 'https://api-${endpoint}.gptbots.ai/v1/bot/doc/text/add' \
-H 'Authorization: Bearer ${API Key}' \
-H 'Content-Type: application/json' \
-d '{
    "knowledge_base_id": "67457fea6f658672d6482542",
    "embedding_model_version_id": "6620f06262390a0be1411c5d",
    "chunk_token": 700,
    "splitter": "\n",
    "files": [
        {
            "file_url": "https://www.gptbots.ai/docs/article_1.pdf",
            "file_base64": "SGVsbG8sIEJhc2U2NCBFbmNvZGluZyE=",
            "source_url": "https://www.gptbots.ai/docs/article_1.pdf",
            "file_name": "article_1.pdf"
        }
    ]
}'

This code block in the floating window

Request Header

Field	Type	Description
Authorization	Bearer ${API Key}	Use Authorization: Bearer ${API Key} for authentication. Get API key from API Key page.
Content-Type	application/json	Data type, set to `application/json`.

Request Parameters

Field	Type	Required	Description
knowledge_base_id	String	No	The target knowledge base to which the document is added. If not filled, it defaults to the "Default" knowledge base.
files	Array<Object>	Yes	List of documents to be added. Supports adding up to 20 documents simultaneously.
file_url	String	No	The URL of the document to be added. Supported formats: pdf/txt/md/doc/docx. Maximum size for PDF is 30MB, and for other formats, it is 10MB. Note: Either URL or base64 must be provided. If both are provided, base64 takes precedence.
file_base64	String	No	The base64 of the document to be added. Supported formats: pdf/txt/md/doc/docx. Maximum size for PDF is 30MB, and for other formats, it is 10MB. Note: Either URL or base64 must be provided. If both are provided, base64 takes precedence.
file_name	String	Yes	The name of the document to be added. 1-200 characters.
source_url	String	No	The source URL of the document to be added. Must comply with URL format specifications.
embedding_model_version_id	String	No	ID of the embedding model version used to vectorize this batch of documents. If omitted, the system default embedding model is used. Obtain a valid value from the `modelId` field returned by Get Model List under the `EMBEDDING` group.
chunk_token	Integer	No	Maximum number of tokens per knowledge chunk during segmentation. Default value is 600. Valid range is 1-1000. Note: Either `chunk_token` or `splitter` must be specified. When both are specified, the splitter takes precedence. If neither is provided, the request still succeeds, but the document fails during vectorization (status `FAIL_STORE`), so always provide at least one.
splitter	String	No	Delimiter used for text chunking. Default is empty. Can use "custom string" as delimiter. Note: Either `chunk_token` or `splitter` must be specified. When both are specified, the splitter takes precedence. If neither is provided, the request still succeeds, but the document fails during vectorization (status `FAIL_STORE`), so always provide at least one.

Response

Response Example

{ "doc": [ { "doc_id": "xxxxxx", "doc_name": "test_1.txt" }, { "doc_id": "xxxxxx", "doc_name": "test_2.pdf" } ], "failed": [ "file_1", "file_2" ] }

                      
                      {
    "doc": [
        {
            "doc_id": "xxxxxx",
            "doc_name": "test_1.txt"
        },
        {
            "doc_id": "xxxxxx",
            "doc_name": "test_2.pdf"
        }
    ],
    "failed": [
        "file_1",
        "file_2"
    ]
}

This code block in the floating window

Success Response

Field	Type	Description
doc	Array<Object>	List of added documents.
doc_id	String	ID of the added document.
doc_name	String	Name of the added document.
failed	Array<Object>	List of failed added document names.

Failure Response

Field	Type	Description
code	Int	Error code.
message	String	Error details.

Error Codes

Code	Message
40000	Parameter error
40000	The embedding_model_version_id does not exist: the given model version ID is not found in the platform model catalog
40000	The embedding_model_version_id is not an embedding model: the given ID refers to a non-embedding model (for example, a chat model `modelId`)
40000	The embedding_model_version_id is not an available embedding model: the provider of this embedding model does not yet support being specified via the API
40127	Developer authentication failed
20059	Agent deleted

Note: If embedding_model_version_id fails validation, no documents in the batch are created and none appear in the failed list; the error response above is returned directly.