Seeds / Upload

Upload Seeds

Upload documents and text content to create searchable seeds. Accepts PDF, Word, Excel, PowerPoint, images, and plain text. Content is automatically processed: text extracted, split into ~300 token chunks, and embedded using Jina Embeddings v4.

POST {baseAPIUrl}/seeds

Authentication

Required. Include your API key in the Authorization header:

Authorization: Bearer YOUR_API_KEY

Query Parameters

ParameterTypeRequiredDescription
appIdstringRequiredApplication identifier
externalUserIdstringRequiredExternal user identifier

Request Body

Content-Type: multipart/form-data

FieldTypeDescription
filesFile[]Binary files (max 10, 100MB total)
fileTypesstring[]Type for each file
fileSourcesstring[]Origin for each file
fileTitlesstring[]Display name for each file
textstring[]Text content as seeds
textTypesstring[]Type for each text
textSourcesstring[]Origin for each text
textTitlesstring[]Display name for each text
bundleIdsstring[]Bundle UUIDs to associate

Supported File Types

Use these values for the fileTypes array:

TypeExtensionsDescription
pdf.pdfPDF documents
word.doc, .docxMicrosoft Word documents
excel.xls, .xlsxMicrosoft Excel spreadsheets
powerpoint.ppt, .pptxMicrosoft PowerPoint presentations
image.jpg, .jpeg, .png, .gif, .webpImages (OCR + visual embedding)
markdown.mdMarkdown files
csv.csvComma-separated values
json.jsonJSON files
html.html, .htmHTML documents
txt.txtPlain text files

Supported Text Types

Use these values for the textTypes array:

TypeDescription
markdownMarkdown-formatted text content
claude_chatExported Claude conversation history
gpt_chatExported ChatGPT conversation history
plainPlain text content
htmlHTML content (will be converted to markdown)
codeSource code snippets

Upload Limits

  • Max files per request: 10 files
  • Max total size: 100MB per request
  • Max file size: 50MB per individual file
  • Max text items: 50 text items per request
  • Chunk size: ~300 tokens with 50 token overlap

Processing Pipeline

1

Upload

Files uploaded to temporary storage

2

Extract

Background job extracts text/images

3

Convert

Content converted to markdown

4

Chunk

Chunked with dynamic token-based splitting

5

Embed

Embeddings generated (Jina Embeddings v4)

6

Store

Stored in database + cloud storage

Typical processing time: 5-30 seconds depending on file size and type.

Try It

Query Parameters

Application identifier

External user identifier for data isolation

Request Body

Binary files to upload (max 10 files, 100MB total)

Type for each file (e.g., "pdf", "word", "image")

Origin identifier for each file

Display name for each file

Text content to upload as seeds

Type for each text item

Origin identifier for each text item

Display name for each text item

Bundle UUIDs to associate with uploaded seeds