Getting Started with Files & Drive

Navigate: Files & Drive → Spaces → Create Space

Space configuration:

```typescript // Space = shared drive (like Google Drive team drive) { name: "Marketing Team", description: "Shared files for the marketing department", storageQuotaBytes: 10737418240, // 10GB limit storageBackend: "r2", // or "s3", "gcs", "azure", "local" isPublic: false, // Private to members members: [ { userId: "usr_abc", role: "admin" }, // Can manage members + settings { userId: "usr_xyz", role: "editor" }, // Can upload/edit/delete ], }

// Storage backend config (in Files Settings): { backend: "r2", bucket: "aeion-files", accountId: "your-cloudflare-account-id", accessKeyId: "...", secretAccessKey: "...", // Encrypted at rest } ```

Permission roles:

• Admin: Manage members, settings, storage quota, retention policies
• Editor: Upload, edit, delete, share files
• Viewer: Read-only access, can download

Space hierarchy:

Marketing Team (space) ├── Campaigns (folder) │ ├── Q1-2026 (folder) │ │ ├── campaign-brief.pdf │ │ └── budget.xlsx │ └── assets (folder) ├── Brand Guidelines (folder) │ └── logo-pack.zip └── Reports (folder)

Navigate: Files & Drive → Settings → Storage

Local disk (development/default):

```bash # No configuration needed — files stored in /var/aeion/files # Works out of the box for development

LOCAL_STORAGE_PATH=/var/aeion/files ```

Amazon S3:

bash S3_BUCKET=my-files S3_REGION=us-east-1 AWS_ACCESS_KEY_ID=AKIA... AWS_SECRET_ACCESS_KEY=... # Optional: custom endpoint for S3-compatible (MinIO) S3_ENDPOINT=https://s3.us-east-1.amazonaws.com

Cloudflare R2:

bash R2_BUCKET=my-files R2_ACCOUNT_ID=your-cloudflare-account-id R2_ACCESS_KEY_ID=... R2_SECRET_ACCESS_KEY=... # R2 uses S3-compatible API — same client as S3 R2_ENDPOINT=https://xxx.r2.cloudflarestorage.com

Google Cloud Storage:

bash GCS_BUCKET=my-files GCS_PROJECT_ID=my-project GCS_KEY_FILE=/path/to/service-account.json # Or use ADC: gcloud auth application-default login

Azure Blob:

bash AZURE_STORAGE_ACCOUNT=myfiles AZURE_STORAGE_KEY=... AZURE_CONTAINER=aeion-files AZURE_STORAGE_ENDPOINT=https://myfiles.blob.core.windows.net

Each space picks one storage backend at creation time. Multi-backend hot/cold tiering with age-based routing rules isn't available yet — if you need it, talk to your account team.

Navigate: Files & Drive → [Space] → Upload

Simple upload (small files < 100MB):

bash # Via API curl -X POST https://api.your-domain.com/files/nodes \ -H "Authorization: Bearer <token>" \ -F "spaceId=spa_abc" \ -F "parentId=nod_xyz" \ -F "file=@report.pdf"

Chunked upload (large files, resumable):

```typescript // Step 1: Initiate const { uploadId, chunkSize, totalChunks } = await initiateUpload({ spaceId: "spa_abc", parentId: "nod_xyz", filename: "video.mp4", size: 524288000, // 500MB mimeType: "video/mp4", });

// Step 2: Upload each chunk const chunkSize = 5242880; // 5MB for (let i = 0; i < totalChunks; i++) { const chunk = fileSlice(i chunkSize, (i + 1) chunkSize); await uploadChunk(uploadId, i, chunk, { onProgress: ({ loaded, total }) => console.log(${loaded}/${total}), }); }

// Step 3: Complete const file = await completeUpload(uploadId); // → Assembles chunks → verifies SHA-256 → stores → version created ```

AI auto-tagging (automatic on upload):

```typescript // Default: AI processing runs on every upload // Disable for specific files: { skipAIProcessing: true, // Privacy-sensitive or large binary files }

// AI results appear in file metadata: { metadata: { aiTags: ["invoice", "Q1-2026", "Acme Corp"], summary: "Invoice for consulting services, $15,000", extractedText: "...", // plus a semantic embedding used for concept-based search } } ```

Navigate: Files & Drive → Settings → WebDAV

The WebDAV panel lets you enable the feature for a space and see its connection status. Full protocol serving — the part that lets Windows/macOS/Linux actually mount a space as a network drive (GET/PUT/MKCOL/DELETE/MOVE/COPY/PROPFIND/LOCK) — isn't wired up in this release, so drive-letter mounting doesn't work yet even with the toggle on.

If your workflow needs file access today, use the browser UI or the upload/download API from Step 3. Ask your account team for the WebDAV rollout timeline if drive mounting is a requirement for your team.

Navigate: Files & Drive → Settings → WOPI

Connect Collabora Online:

```bash # Option 1: Self-hosted Collabora COLLABORA_URL=https://collabora.your-domain.com COLLABORA_SECRET=your-secret

# Option 2: ONLYOFFICE ONLYOFFICE_URL=https://onlyoffice.your-domain.com ONLYOFFICE_SECRET=your-secret ```

WOPI flow:

```typescript // 1. User opens a .docx file in Files & Drive // 2. Clicks "Edit in Browser" (WOPI button) // 3. Files module creates WOPI session: { fileId: "nod_abc", wopiServerUrl: "https://collabora.your-domain.com", accessToken: "wopi_token_xxx", expiresAt: "2026-05-11T11:00:00Z", }

// 4. Browser loads Collabora/ONLYOFFICE iframe with access token // 5. WOPI server fetches file content from Files module // 6. User edits in browser // 7. WOPI server saves back to Files module on close

// Supported formats: .docx, .xlsx, .pptx ```

Current status: file locking (open-for-edit / release) is fully wired. Saving edits back to Aeion Files on close isn't persisted yet in this release — treat browser-based editing as view/annotate for now, and keep using upload/download for anything you need saved.

Navigate: Files & Drive → [Space] → Settings → Retention

Retention policy example:

```typescript // Policy for a compliance-required space { name: "Financial Records Retention", rules: [ { retentionDays: 2555, action: "archive" }, // 7 years (SOX compliance) { retentionDays: 365, action: "delete" }, // auto-cleanup after 1 year ], }

// Policy for a project space (shorter retention) { name: "Project Files", rules: [ { retentionDays: 90, action: "delete" }, ], } ```

Current enforcement: per-rule enforcement of the policy above isn't wired up yet. Today, a separate daily cron prunes old versions using a simple global version-count cap, not your configured per-rule policy — treat the Retention tab as your policy of record for audits, but verify actual cleanup behavior directly until per-rule enforcement ships.

Version labels: Mark versions as "draft", "final", "reviewed" for manual protection. Labeled versions are exempt from auto-cleanup unless explicitly included.

Navigate: Files & Drive → Search

Full-text search:

```bash # Search file names and content POST /api/files/search?q=invoice+Acme # → Matches: "invoice" AND "Acme" in filenames, extracted text, tags

# Filters ?q=invoice&mimeType=application/pdf&spaceId=spa_abc&createdAfter=2026-01-01

# Full query syntax ?q=(invoice OR contract) AND tag:financial AND size:>1MB ```

Semantic search (AI-powered):

typescript // "Find files similar to this contract" const similar = await aiService.findSimilarFiles({ fileId: "nod_abc", limit: 10, }); // → Ranked by embedding similarity, not just keyword match

OCR for scanned documents: Images and PDFs without text layers go through OCR automatically. Scanned contracts, receipts, and documents become searchable.