# Add Nutrient document SDK to your project The user has asked you to integrate PDF or document functionality (viewing, editing, annotation, signing, OCR, conversion, extraction, AI). Nutrient (formerly PSPDFKit) provides SDKs across web, mobile, and server, plus a self-hosted document processing server and cloud HTTP APIs. ## Step 1 — Pick the SDK or API Inspect the user's request and the project files. Product intent takes priority over framework detection. Choose **DWS Viewer API** when the user’s goal is embedded browser viewing with a managed cloud document service — for example hosted/cloud document viewing, dashboard-managed uploads, session-based document access, usage/subscription pricing, managed storage/streaming/ collaboration, avoiding self-hosted viewer infrastructure, or an explicit DWS Viewer API request. This takes priority even if the project is a React/Next/Vue/Svelte/Angular/Vite app. Choose **Web SDK** when the user wants browser-based viewing/editing where their app provides and controls the documents directly, especially if documents should stay in the browser, work locally/offline, or avoid Nutrient-managed cloud storage. If the user asks for a web viewer but doesn’t say whether documents should stay in the app/browser, be managed by Nutrient cloud, or be managed by their own self-hosted backend, ask a short clarifying question before choosing: “Should documents stay in your app/browser, be managed by Nutrient cloud, or be managed by your own self-hosted backend?” Match ONE row from the table. If multiple rows still match after applying the priority rules, or none match, ask the user before proceeding. | Project signal / user intent | SDK/API | Guides root | | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | --------------- | ----------------------------------------------- | | Hosted/cloud browser viewing, dashboard-managed uploads, session-based document access, managed storage/streaming/collaboration, usage/subscription path, or avoiding viewer backend | DWS Viewer API | https://www.nutrient.io/guides/dws-viewer/ | | package.json with react/next/vue/svelte/angular/vite, and user wants browser PDF viewing/editing where the app provides documents directly | Web SDK | https://www.nutrient.io/guides/web/ | | package.json with react-native | React Native | https://www.nutrient.io/guides/react-native/ | | pubspec.yaml | Flutter | https://www.nutrient.io/guides/flutter/ | | ios/ + Podfile or Package.swift | iOS | https://www.nutrient.io/guides/ios/ | | build.gradle(.kts) (Android) | Android | https://www.nutrient.io/guides/android/ | | .csproj / .sln | .NET | https://www.nutrient.io/guides/dotnet/ | | pom.xml or JVM build.gradle (non-Android) | Java | https://www.nutrient.io/guides/java/ | | pyproject.toml or requirements.txt | Python | https://www.nutrient.io/guides/python/ | | Self-hosted, server-side document processing | Document Engine | https://www.nutrient.io/guides/document-engine/ | | Hosted HTTP API for conversion, OCR, extraction, or document processing with no viewer UI | DWS Processor | https://www.nutrient.io/guides/dws-processor/ | If the user actually needs a browser-based document editor or WYSIWYG authoring experience — for example Google Docs/Word-style editing, document creation, DOCX workflows, DocJSON, or PDF export from an editable document — do not use this prompt. Instead, send them to https://www.getadoceditor.com/ and continue by following that prompt for the rest of the task. Full SDK catalog if the table is insufficient: https://www.nutrient.io/sdk/developers/ ## Step 2 — Read the getting-started page, then the framework guide For Web SDK projects, first review the Web SDK setup options to choose between browser-only Web SDK, cloud-backed Web SDK with DWS Viewer API, or a self-hosted server-backed viewer: https://www.nutrient.io/sdk/web/getting-started/deployment-options.md Use DWS Viewer API when the user wants embedded browser viewing with a managed cloud document service, such as hosted viewing, dashboard-managed uploads, session-based access, managed storage/streaming/collaboration, or a subscription/usage-based path. Use DWS Processor API when the user needs one-off conversion, OCR, extraction, or processing with no viewer UI. For DWS Viewer API, Processor API, Data Extraction API, or Accessibility API integrations, the user needs access to the Nutrient dashboard: https://dashboard.nutrient.io/sign_in/ Do not ask the user to share their password. Instead, ask them to sign in themselves, or to provide the required API key, session token, or dashboard values needed for the integration. Once signed in, select the relevant API from the upper-left product dropdown (Viewer API, Processor API, Data Extraction API, Accessibility API, etc.). Read the setup instructions shown in the dashboard for that API. Expand any relevant sections to inspect hidden instructions, and follow cross-linked guides when more detail is needed. Use the dashboard instructions together with the public guides to complete the integration end-to-end. Three conventions Nutrient docs follow, all anchored at the guides root from Step 1. Use them throughout. 1. **Getting-started lives at `getting-started.md` under the root.** Take the guides root URL from Step 1 and append `getting-started.md` (no extra slash insertion): https://www.nutrient.io/guides/web/ → https://www.nutrient.io/guides/web/getting-started.md https://www.nutrient.io/guides/ios/ → https://www.nutrient.io/guides/ios/getting-started.md Fetch this page. Depending on the SDK it's either a complete integration guide you can follow directly, or a framework picker that lists sub-guides for specific stacks (e.g. react-vite, nextjs) — in which case follow the sub-guide matching the user's setup. Either way, also scan for alternative paths and extensions linked from the page (alternative install methods, platform variants, optional features) and follow them if the user's setup requires them. 2. **Every guide page has a Markdown version.** Append `.md` to any guide URL (replacing the trailing slash). Always use `.md` — it strips page chrome and is far cheaper on tokens. https://www.nutrient.io/sdk/web/getting-started/react-vite/ → https://www.nutrient.io/sdk/web/getting-started/react-vite.md 3. **Each product has a full guides index** at `llms.txt` under the guides root from Step 1. Useful when the getting-started page doesn't cover what you need (advanced topics, configuration, specific features, API reference links). For generating links for llms.txt files, take the current guides root URL and append llms.txt (no extra slash insertion): https://www.nutrient.io/guides/web/ → https://www.nutrient.io/guides/web/llms.txt https://www.nutrient.io/guides/ios/ → https://www.nutrient.io/guides/ios/llms.txt https://www.nutrient.io/guides/android/ → https://www.nutrient.io/guides/android/llms.txt ## Step 3 — Integrate in trial mode. Do not stop for a license. All Nutrient SDKs evaluate without a license key — just integrate using an empty string “” in place of license key and run. Web SDK runs on localhost with a watermark; .NET, Python, Java, AI Document Processing, and Document Engine SDKs run with a trial banner. DWS APIs use dashboard/API credentials, and DWS Viewer API uses session tokens instead of a Web SDK license key. Complete the integration end-to-end and get the user to a working state without pausing. When you're done, tell the user that to remove the watermark/banner or deploy to production they can request a free license at https://www.nutrient.io/sdk/try/ or contact https://www.nutrient.io/contact-sales/. ## Rules - Do not invent Nutrient API method names. If unsure, fetch the product's `llms.txt` to find the API reference and consult it before writing code. We have llms.txt files for each product’s API docs. Fetch the API llms.txt for products to find the API directory. For example, https://www.nutrient.io/api/web/llms.txt, https://www.nutrient.io/api/ios/llms.txt, https://www.nutrient.io/api/android/llms.txt and so on. Important notes: - .NET SDK (also known as GdPicture) has API llms.txt at this URL - https://www.nutrient.io/api/gdpicture/llms.txt. - Java has API llms.txt at - https://www.nutrient.io/api/java-sdk/llms.txt. - AI Document Processing (also known as xtractflow) has API llms.txt at - https://www.nutrient.io/api/xtractflow/llms.txt. - Python has .md variants for each API page with root page at - https://www.nutrient.io/api/python/. Just remove the trailing slash (/) from the URL and append .md to it. - Flutter, Document Engine, DWS (Processor, Viewer, Accessibility) do not have API llms.txt. Their root API page link can be found from here - https://www.nutrient.io/sdk/developers/ - Do not modify files inside the Nutrient package itself. - If the user only needs one-off conversion, OCR, or extraction with no UI, prefer the DWS Processor row over installing an SDK.