Skip to main content
The Documents API provides endpoints for managing documents in the vault, including upload, processing, tagging, and retrieval.

​
Query Endpoints

​
get

Retrieve a paginated list of documents.
​
q
string
Search query for document name or content
​
tag
string
Filter by tag name
​
start
string
Start date filter (ISO 8601)
​
end
string
End date filter (ISO 8601)
​
sort
array
Array of sort objects
​
limit
number
Number of results per page
import { trpc } from '@/lib/trpc';

const { data, meta } = await trpc.documents.get.query({
  tag: 'receipts',
  start: '2024-01-01',
  end: '2024-12-31',
});

​
getById

Retrieve a specific document by ID or file path.
​
id
string
Document UUID
​
filePath
array
Array of path tokens (alternative to id)
​
document
object
const document = await trpc.documents.getById.query({
  id: 'document-uuid',
});

// Or by file path
const document = await trpc.documents.getById.query({
  filePath: ['vault', 'team-id', 'receipt.pdf'],
});

​
getRelatedDocuments

Get documents related to a specific document (based on tags, date, etc.).
​
id
string
required
Document UUID
​
pageSize
number
Number of results to return
​
documents
array
Array of related document objects
const related = await trpc.documents.getRelatedDocuments.query({
  id: 'document-uuid',
  pageSize: 5,
});

​
checkAttachments

Check if a document is attached to any transactions or invoices.
​
id
string
required
Document UUID
​
data
object
const { hasAttachments, transactionCount } = await trpc.documents.checkAttachments.query({
  id: 'document-uuid',
});

if (hasAttachments) {
  console.log(`Document is attached to ${transactionCount} transactions`);
}

​
Mutation Endpoints

​
delete

Delete a document from the vault.
​
id
string
required
Document UUID
​
document
object
The deleted document object
await trpc.documents.delete.mutate({
  id: 'document-uuid',
});
Deleting a document also removes it from storage. This action cannot be undone.

​
processDocument

Trigger processing for uploaded documents (OCR, text extraction).
​
documents
array
required
Array of document objects to process
type DocumentToProcess = {
  filePath: string[];  // Path tokens in storage
  mimetype: string;    // MIME type (e.g., 'application/pdf')
};
​
data
object
const result = await trpc.documents.processDocument.mutate([
  {
    filePath: ['vault', 'team-id', 'invoice-001.pdf'],
    mimetype: 'application/pdf',
  },
  {
    filePath: ['vault', 'team-id', 'receipt-002.jpg'],
    mimetype: 'image/jpeg',
  },
]);

console.log('Processing jobs:', result.jobs);
Supported file types for processing: PDF, images (JPEG, PNG), and text files. Other file types will be marked as completed without processing.

​
reprocessDocument

Reprocess a document that failed or needs updating.
​
id
string
required
Document UUID
​
data
object
const result = await trpc.documents.reprocessDocument.mutate({
  id: 'document-uuid',
});

if (result.success) {
  console.log('Reprocessing job ID:', result.jobId);
}

​
Signed URLs

​
signedUrl

Generate a temporary signed URL for downloading a document.
​
filePath
array
required
Array of path tokens
​
expireIn
number
Expiration time in seconds (default: 60)
​
signedUrl
string
Temporary URL for downloading the file
const { signedUrl } = await trpc.documents.signedUrl.mutate({
  filePath: ['vault', 'team-id', 'receipt.pdf'],
  expireIn: 300, // 5 minutes
});

// Use the signed URL to download or display the file
window.open(signedUrl);

​
signedUrls

Generate multiple signed URLs at once.
​
filePaths
array
required
Array of file path arrays
​
urls
array
Array of signed URLs (same order as input)
const urls = await trpc.documents.signedUrls.mutate([
  ['vault', 'team-id', 'file1.pdf'],
  ['vault', 'team-id', 'file2.jpg'],
  ['vault', 'team-id', 'file3.png'],
]);

// urls is an array of signed URL strings
urls.forEach((url, index) => {
  console.log(`File ${index + 1}: ${url}`);
});

​
Document Processing

The document processing system automatically extracts text and metadata from uploaded files.

​
Processing Flow

  1. Upload - Files are uploaded to Supabase storage
  2. Queue - Call processDocument to queue processing jobs
  3. Process - Background workers extract text using OCR/parsers
  4. Complete - Document status changes to “completed”

​
Supported File Types

  • PDF - Text extraction and OCR for scanned PDFs
  • Images - JPEG, PNG (OCR applied)
  • Text - Plain text files

​
Example: Upload and Process

// 1. Upload file to storage
const filePath = await uploadToStorage(file);

// 2. Queue processing
const result = await trpc.documents.processDocument.mutate([{
  filePath,
  mimetype: file.type,
}]);

// 3. Monitor status
const document = await trpc.documents.getById.query({ filePath });
console.log('Processing status:', document.processingStatus);

​
Tagging

Documents can be organized using tags through the tRPC API’s documentTags and documentTagAssignments routers.

Build docs developers (and LLMs) love

Get started for freeTalk to us