npm SDK File Loading

Once you have a resource ID, openFile is the one method to remember. Tell it which viewer you are using and it will return the matching data:

import { openFile } from '@artosai/sdk'

const result = await openFile({
    resourceId,
    accessToken,
    baseUrl,
    viewer: 'apryse' // or 'onlyoffice'
})

Here are the two possible responses:

// viewer: 'apryse'
{
    viewer: 'apryse',
    blob: Blob,
    contentType: string,
    fileName: string | null,
    resourceId: string,
    sourceFileUrl?: string
}

// viewer: 'onlyoffice'
{
    viewer: 'onlyoffice',
    config: object | null,
    token: string | null,
    resourceId: string,
    _raw: object
}

The two responses are deliberately different. The Apryse response has the file bytes; the OnlyOffice response has the editor inputs. Check result.viewer before reading the viewer-specific fields.

Open the returned file

openFile retrieves and authorizes the resource, but it does not create a viewer component. After an Apryse call, hand the returned Blob to Apryse or create a temporary browser URL:

const result = await openFile({
    resourceId,
    viewer: 'apryse',
    accessToken,
    baseUrl
})

if (result.viewer !== 'apryse') {
    throw new Error('Expected an Apryse file response')
}

// Apryse WebViewer example:
import WebViewer from '@pdftron/webviewer'

const viewerElement = document.getElementById('apryse-viewer')
const instance = await WebViewer({ path: '/webviewer' }, viewerElement)
await instance.UI.loadDocument(result.blob, {
    filename: result.fileName || 'document'
})

For a quick browser preview, use an object URL:

const previewUrl = URL.createObjectURL(result.blob)
window.open(previewUrl, '_blank')

// Revoke it when the preview is closed or replaced—not immediately after
// window.open(), because the new page may not have loaded the Blob yet.

To use OnlyOffice, load its editor API script and add an editor container:

<div id="onlyoffice-editor"></div>
<script src="https://YOUR-ONLYOFFICE-HOST/web-apps/apps/api/documents/api.js"></script>

Then pass the returned editor inputs to the OnlyOffice editor—not the Blob fields:

const result = await openFile({
    resourceId,
    viewer: 'onlyoffice',
    mode: 'edit',
    accessToken,
    baseUrl
})

if (result.viewer !== 'onlyoffice') {
    throw new Error('Expected an OnlyOffice response')
}

// The OnlyOffice API script must already be loaded and this element must exist:
// <div id="onlyoffice-editor"></div>
const editor = new DocsAPI.DocEditor('onlyoffice-editor', {
    ...result.config,
    token: result.token
})

The viewer libraries are still part of your application. The SDK supplies the authorized file or editor configuration; Apryse and OnlyOffice render it.

Documents

For a regular document preview or download, use the document ID as the resource ID:

const result = await openFile({
    resourceId: documentId,
    viewer: 'apryse',
    accessToken,
    baseUrl
})

if (result.viewer === 'apryse') {
    const previewUrl = URL.createObjectURL(result.blob)
    // Pass previewUrl to your viewer, then revoke it when replaced.
}

Templates

Templates work the same way, but you first ask Artos for the template resource ID:

import { getTemplateResource, openFile } from '@artosai/sdk'

const template = await getTemplateResource({
    documentId,
    accessToken,
    baseUrl
})

if (!template.templateResourceId) {
    throw new Error('This document has no template file')
}

const result = await openFile({
    resourceId: template.templateResourceId,
    viewer: 'apryse',
    accessToken,
    baseUrl
})

Source files

Source files need one extra lookup. Resolve the source reference to a resource ID, then pass that ID to openFile:

import {
    getDocumentSources,
    getDocumentSourceResource,
    getSectionFromText,
    openFile
} from '@artosai/sdk'

const section = await getSectionFromText({
    documentId,
    highlightedText,
    accessToken,
    baseUrl
})

const { sources } = await getDocumentSources({
    documentId,
    sectionId: section.sectionId,
    accessToken,
    baseUrl
})

if (sources.length === 0 || !sources[0].documentSourceId) {
    throw new Error('No source file was found for this section')
}

const source = await getDocumentSourceResource({
    documentSourceId: sources[0].documentSourceId,
    accessToken,
    baseUrl
})

const result = await openFile({
    resourceId: source.resourceId,
    viewer: 'apryse',
    accessToken,
    baseUrl
})

For sourcefile-... resources, openFile uses authenticated POST /source-file. Documents, templates, and generated outputs use the document-session flow. In both cases, you pass the resource ID; you do not construct an S3 URL or append a session token.

OnlyOffice

To open a resource in OnlyOffice, use the same method with viewer: 'onlyoffice':

const result = await openFile({
    resourceId,
    viewer: 'onlyoffice',
    mode: 'edit', // or 'readonly'
    accessToken,
    baseUrl
})

if (result.viewer === 'onlyoffice') {
    new DocsAPI.DocEditor('onlyoffice-editor', {
        ...result.config,
        token: result.token
    })
}

edit requests an editable editor. readonly disables editing and saving. Source files, templates, generated outputs, and duplicate outputs may be forced read-only by the backend regardless of the requested mode. For editable documents, OnlyOffice sends saves through the callback in the generated config. Keep config and token from the same response, leave the config unchanged, and request a new result when reopening an expired editor session.

Getting Started

API Reference

Core Concepts

Single-Tenant

Cookbooks

NPM SDK Reference

Integrations

Deployment

npm SDK File Loading

npm SDK File Loading

Open the returned file

Documents

Templates

Source files

OnlyOffice

​npm SDK File Loading

​Open the returned file

​Documents

​Templates

​Source files

​OnlyOffice

npm SDK File Loading

Open the returned file

Documents

Templates

Source files

OnlyOffice