Skip to content

Papermark is the open-source DocSend alternative and secure data rooms with built-in analytics and custom domains.

feat(ee): add ai processing for various document types

Marc Seitz08bbc630de35
diff --git a/app/(ee)/api/ai/store/teams/[teamId]/datarooms/[dataroomId]/route.ts b/app/(ee)/api/ai/store/teams/[teamId]/datarooms/[dataroomId]/route.ts
index cffacb26..69170dee 100644
--- a/app/(ee)/api/ai/store/teams/[teamId]/datarooms/[dataroomId]/route.ts
+++ b/app/(ee)/api/ai/store/teams/[teamId]/datarooms/[dataroomId]/route.ts
@@ -1,8 +1,11 @@
import { NextRequest, NextResponse } from "next/server";
-import { processDocumentForVectorStore } from "@/ee/features/ai/lib/file-processing/process-document-for-vector-store";
+import {
+ SUPPORTED_AI_CONTENT_TYPES,
+ addFileToVectorStoreTask,
+ processDocumentForAITask,
+} from "@/ee/features/ai/lib/trigger";
import { createDataroomVectorStore } from "@/ee/features/ai/lib/vector-stores/create-dataroom-vector-store";
-import { addFileToVectorStore } from "@/ee/features/ai/lib/vector-stores/upload-file-to-vector-store";
import { authOptions } from "@/pages/api/auth/[...nextauth]";
import { getServerSession } from "next-auth";
@@ -13,6 +16,7 @@ import { CustomUser } from "@/lib/types";
/**
* POST /api/ai/store/teams/[teamId]/datarooms/[dataroomId]
* Index all documents in a dataroom into its vector store
+ * Returns batch runIds for status tracking via polling
*/
export async function POST(
req: NextRequest,
@@ -66,10 +70,20 @@ export async function POST(
documents: {
include: {
document: {
- include: {
+ select: {
+ id: true,
+ name: true,
versions: {
where: { isPrimary: true },
take: 1,
+ select: {
+ id: true,
+ fileId: true,
+ contentType: true,
+ storageType: true,
+ originalFile: true,
+ file: true,
+ },
},
},
},
@@ -117,10 +131,16 @@ export async function POST(
});
}
- // Process and upload documents
- let filesIndexed = 0;
+ // Track triggered runs for status polling
+ const triggeredRuns: Array<{
+ documentId: string;
+ documentName: string;
+ runId: string;
+ }> = [];
+ const skippedDocuments: string[] = [];
const errors: string[] = [];
+ // Trigger processing for each document
for (const dataroomDoc of dataroom.documents) {
const document = dataroomDoc.document;
const primaryVersion = document.versions[0];
@@ -130,41 +150,29 @@ export async function POST(
continue;
}
- // Skip if already indexed
+ // Skip if already indexed in this dataroom
if (dataroomDoc.vectorStoreFileId) {
- filesIndexed++;
+ skippedDocuments.push(document.name);
continue;
}
- try {
- // Check if document type is supported
- const supportedTypes = ["application/pdf"];
- if (!supportedTypes.includes(primaryVersion.contentType || "")) {
- errors.push(
- `Unsupported file type for document: ${document.name} (${primaryVersion.contentType})`,
- );
- continue;
- }
-
- // get version's fileId
- let fileId = primaryVersion.fileId;
-
- if (!fileId) {
- const { fileId: newFileId } = await processDocumentForVectorStore(
- primaryVersion.file,
- primaryVersion.storageType,
- );
-
- // Update document version with file ID
- await prisma.documentVersion.update({
- where: { id: primaryVersion.id },
- data: { fileId: newFileId },
- });
+ // Check if document type is supported
+ const contentType = primaryVersion.contentType || "";
+ if (!SUPPORTED_AI_CONTENT_TYPES.includes(contentType)) {
+ errors.push(
+ `Unsupported file type for document: ${document.name} (${contentType})`,
+ );
+ continue;
+ }
- fileId = newFileId;
- }
+ try {
+ // Determine file path
+ const filePath =
+ primaryVersion.originalFile && contentType !== "application/pdf"
+ ? primaryVersion.originalFile
+ : primaryVersion.file;
- const metadata = {
+ const fileMetadata = {
teamId: dataroom.teamId,
documentId: document.id,
documentName: document.name,
@@ -174,33 +182,65 @@ export async function POST(
dataroomFolderId: dataroomDoc.folderId || "root",
};
- // Add to vector store
- const vectorStoreFileId = await addFileToVectorStore({
- vectorStoreId,
- fileId,
- metadata,
- });
+ let handle;
- // Update dataroom document with vector store file ID
- await prisma.dataroomDocument.update({
- where: { id: dataroomDoc.id },
- data: { vectorStoreFileId },
- });
+ // If document already has fileId, just add to vector store
+ if (primaryVersion.fileId) {
+ handle = await addFileToVectorStoreTask.trigger({
+ fileId: primaryVersion.fileId,
+ vectorStoreId,
+ metadata: fileMetadata,
+ });
+ } else {
+ // Trigger full processing
+ handle = await processDocumentForAITask.trigger(
+ {
+ documentId: document.id,
+ documentVersionId: primaryVersion.id,
+ teamId: dataroom.teamId,
+ vectorStoreId,
+ documentName: document.name,
+ filePath,
+ storageType: primaryVersion.storageType,
+ contentType,
+ metadata: fileMetadata,
+ },
+ {
+ idempotencyKey: `ai-index-dataroom-${dataroom.id}-${primaryVersion.id}`,
+ tags: [
+ `team_${dataroom.teamId}`,
+ `dataroom_${dataroom.id}`,
+ `document_${document.id}`,
+ `version_${primaryVersion.id}`,
+ ],
+ },
+ );
+ }
- filesIndexed++;
+ triggeredRuns.push({
+ documentId: document.id,
+ documentName: document.name,
+ runId: handle.id,
+ });
} catch (error) {
- console.error(`Error indexing document ${document.name}:`, error);
+ console.error(
+ `Error triggering processing for ${document.name}:`,
+ error,
+ );
errors.push(
- `Failed to index document: ${document.name} - ${error instanceof Error ? error.message : "Unknown error"}`,
+ `Failed to trigger processing for: ${document.name} - ${error instanceof Error ? error.message : "Unknown error"}`,
);
}
}
return NextResponse.json({
success: true,
- filesIndexed,
- totalDocuments: dataroom.documents.length,
vectorStoreId,
+ totalDocuments: dataroom.documents.length,
+ triggeredCount: triggeredRuns.length,
+ skippedCount: skippedDocuments.length,
+ runs: triggeredRuns,
+ skipped: skippedDocuments.length > 0 ? skippedDocuments : undefined,
errors: errors.length > 0 ? errors : undefined,
});
} catch (error) {
diff --git a/app/(ee)/api/ai/store/teams/[teamId]/documents/[documentId]/route.ts b/app/(ee)/api/ai/store/teams/[teamId]/documents/[documentId]/route.ts
index 1da9f4ad..b5ec38de 100644
--- a/app/(ee)/api/ai/store/teams/[teamId]/documents/[documentId]/route.ts
+++ b/app/(ee)/api/ai/store/teams/[teamId]/documents/[documentId]/route.ts
@@ -1,9 +1,12 @@
import { NextRequest, NextResponse } from "next/server";
-import { processDocumentForVectorStore } from "@/ee/features/ai/lib/file-processing/process-document-for-vector-store";
+import {
+ addFileToVectorStoreTask,
+ processDocumentForAITask,
+ SUPPORTED_AI_CONTENT_TYPES,
+} from "@/ee/features/ai/lib/trigger";
import { createTeamVectorStore } from "@/ee/features/ai/lib/vector-stores/create-team-vector-store";
import { removeFileFromVectorStore } from "@/ee/features/ai/lib/vector-stores/remove-file-from-vector-store";
-import { addFileToVectorStore } from "@/ee/features/ai/lib/vector-stores/upload-file-to-vector-store";
import { authOptions } from "@/pages/api/auth/[...nextauth]";
import { getServerSession } from "next-auth";
@@ -14,6 +17,7 @@ import { CustomUser } from "@/lib/types";
/**
* POST /api/ai/store/teams/[teamId]/documents/[documentId]
* Index a single document into the team vector store
+ * Returns runId for status tracking via polling
*/
export async function POST(
req: NextRequest,
@@ -118,34 +122,23 @@ export async function POST(
}
// Check if document type is supported
- const supportedTypes = ["application/pdf"];
- if (!supportedTypes.includes(primaryVersion.contentType || "")) {
+ const contentType = primaryVersion.contentType || "";
+ if (!SUPPORTED_AI_CONTENT_TYPES.includes(contentType)) {
return NextResponse.json(
{
- error: `Unsupported file type: ${primaryVersion.contentType}. Only PDF files are supported.`,
+ error: `Unsupported file type: ${contentType}. Supported types: PDF, Excel, and images (JPEG, PNG, WebP).`,
},
{ status: 400 },
);
}
- let fileId = primaryVersion.fileId;
+ // Determine file path - use originalFile for non-PDF, file for PDF
+ const filePath =
+ primaryVersion.originalFile && contentType !== "application/pdf"
+ ? primaryVersion.originalFile
+ : primaryVersion.file;
- if (!fileId) {
- const { fileId: newFileId } = await processDocumentForVectorStore(
- primaryVersion.file,
- primaryVersion.storageType,
- );
-
- // Update document version with file ID
- await prisma.documentVersion.update({
- where: { id: primaryVersion.id },
- data: { fileId: newFileId },
- });
-
- fileId = newFileId;
- }
-
- const metadata = {
+ const fileMetadata = {
teamId: document.teamId,
documentId: document.id,
documentName: document.name,
@@ -153,23 +146,49 @@ export async function POST(
folderId: document.folderId || "root",
};
- // Add to vector store
- const vectorStoreFileId = await addFileToVectorStore({
- vectorStoreId,
- fileId,
- metadata,
- });
+ // Check if document already has a fileId - use lightweight task
+ if (primaryVersion.fileId) {
+ // Document already processed, just add to vector store
+ const handle = await addFileToVectorStoreTask.trigger({
+ fileId: primaryVersion.fileId,
+ vectorStoreId,
+ metadata: fileMetadata,
+ });
- // Update document version with vector store file ID
- await prisma.documentVersion.update({
- where: { id: primaryVersion.id },
- data: { vectorStoreFileId },
- });
+ return NextResponse.json({
+ success: true,
+ runId: handle.id,
+ isReprocessing: false,
+ });
+ }
+
+ // Trigger full document processing task
+ const handle = await processDocumentForAITask.trigger(
+ {
+ documentId: document.id,
+ documentVersionId: primaryVersion.id,
+ teamId: document.teamId,
+ vectorStoreId,
+ documentName: document.name,
+ filePath,
+ storageType: primaryVersion.storageType,
+ contentType,
+ metadata: fileMetadata,
+ },
+ {
+ idempotencyKey: `ai-index-${document.teamId}-${primaryVersion.id}`,
+ tags: [
+ `team_${document.teamId}`,
+ `document_${document.id}`,
+ `version_${primaryVersion.id}`,
+ ],
+ },
+ );
return NextResponse.json({
success: true,
- vectorStoreFileId,
- vectorStoreId,
+ runId: handle.id,
+ isReprocessing: true,
});
} catch (error) {
console.error("Error indexing document:", error);
diff --git a/ee/features/ai/lib/models/google.ts b/ee/features/ai/lib/models/google.ts
new file mode 100644
index 00000000..3f5bf16b
--- /dev/null
+++ b/ee/features/ai/lib/models/google.ts
@@ -0,0 +1,7 @@
+import { createVertex } from "@ai-sdk/google-vertex";
+
+const vertex = createVertex({
+ apiKey: process.env.GOOGLE_VERTEX_API_KEY,
+});
+
+export { vertex };
diff --git a/ee/features/ai/lib/trigger/add-file-to-vector-store.ts b/ee/features/ai/lib/trigger/add-file-to-vector-store.ts
new file mode 100644
index 00000000..39a9199a
--- /dev/null
+++ b/ee/features/ai/lib/trigger/add-file-to-vector-store.ts
@@ -0,0 +1,55 @@
+import { logger, metadata, task } from "@trigger.dev/sdk/v3";
+
+import { openai } from "@/ee/features/ai/lib/models/openai";
+
+import type { AddToVectorStorePayload } from "./types";
+
+/**
+ * Add an existing OpenAI file to a vector store
+ * This is used when a document has already been processed and we just need to add it
+ * to a new vector store (e.g., when adding to a dataroom)
+ */
+export const addFileToVectorStoreTask = task({
+ id: "add-file-to-vector-store",
+ retry: { maxAttempts: 3 },
+ queue: {
+ concurrencyLimit: 10,
+ },
+ run: async (
+ payload: AddToVectorStorePayload,
+ ): Promise<{ vectorStoreFileId: string }> => {
+ const { fileId, vectorStoreId, metadata: fileMetadata } = payload;
+
+ logger.info("Adding file to vector store", {
+ fileId,
+ vectorStoreId,
+ documentId: fileMetadata.documentId,
+ });
+
+ metadata.set("status", "indexing");
+ metadata.set("step", "Adding to vector store...");
+ metadata.set("progress", 80);
+
+ // Add file to vector store with metadata
+ const vectorStoreFile = await openai.vectorStores.files.create(
+ vectorStoreId,
+ {
+ file_id: fileId,
+ attributes: fileMetadata,
+ },
+ );
+
+ metadata.set("status", "completed");
+ metadata.set("step", "Indexed successfully");
+ metadata.set("progress", 100);
+ metadata.set("vectorStoreFileId", vectorStoreFile.id);
+
+ logger.info("File added to vector store successfully", {
+ fileId,
+ vectorStoreFileId: vectorStoreFile.id,
+ documentId: fileMetadata.documentId,
+ });
+
+ return { vectorStoreFileId: vectorStoreFile.id };
+ },
+});
diff --git a/ee/features/ai/lib/trigger/index.ts b/ee/features/ai/lib/trigger/index.ts
new file mode 100644
index 00000000..4b318acd
--- /dev/null
+++ b/ee/features/ai/lib/trigger/index.ts
@@ -0,0 +1,9 @@
+// Export all AI processing tasks
+export { addFileToVectorStoreTask } from "./add-file-to-vector-store";
+export { processDocumentForAITask } from "./process-document-for-ai";
+export { processExcelForAITask } from "./process-excel-for-ai";
+export { processImageForAITask } from "./process-image-for-ai";
+export { processPdfForAITask } from "./process-pdf-for-ai";
+
+// Export types
+export * from "./types";
diff --git a/ee/features/ai/lib/trigger/process-document-for-ai.ts b/ee/features/ai/lib/trigger/process-document-for-ai.ts
new file mode 100644
index 00000000..4717ee73
--- /dev/null
+++ b/ee/features/ai/lib/trigger/process-document-for-ai.ts
@@ -0,0 +1,194 @@
+import { logger, metadata, task } from "@trigger.dev/sdk/v3";
+
+import prisma from "@/lib/prisma";
+
+import { addFileToVectorStoreTask } from "./add-file-to-vector-store";
+import { processExcelForAITask } from "./process-excel-for-ai";
+import { processImageForAITask } from "./process-image-for-ai";
+import { processPdfForAITask } from "./process-pdf-for-ai";
+import {
+ EXCEL_CONTENT_TYPES,
+ IMAGE_CONTENT_TYPES,
+ PDF_CONTENT_TYPES,
+ SUPPORTED_AI_CONTENT_TYPES,
+ type ProcessDocumentPayload,
+} from "./types";
+
+/**
+ * Main orchestrator task for processing documents for AI
+ * Routes to appropriate subtask based on content type, then adds to vector store
+ */
+export const processDocumentForAITask = task({
+ id: "process-document-for-ai",
+ retry: { maxAttempts: 3 },
+ queue: {
+ concurrencyLimit: 10,
+ },
+ run: async (
+ payload: ProcessDocumentPayload,
+ ): Promise<{ vectorStoreFileId: string; fileId: string }> => {
+ const {
+ documentId,
+ documentVersionId,
+ teamId,
+ vectorStoreId,
+ documentName,
+ filePath,
+ storageType,
+ contentType,
+ metadata: fileMetadata,
+ } = payload;
+
+ logger.info("Starting document processing for AI", {
+ documentId,
+ documentVersionId,
+ teamId,
+ contentType,
+ });
+
+ // Initialize metadata for real-time tracking
+ metadata.set("status", "pending");
+ metadata.set("documentName", documentName);
+ metadata.set("documentId", documentId);
+ metadata.set("step", "Initializing...");
+ metadata.set("progress", 0);
+
+ // Validate content type
+ if (!SUPPORTED_AI_CONTENT_TYPES.includes(contentType)) {
+ metadata.set("status", "failed");
+ metadata.set("step", "Unsupported file type");
+ metadata.set("error", `Unsupported content type: ${contentType}`);
+ throw new Error(`Unsupported content type: ${contentType}`);
+ }
+
+ // Check if fileId already exists
+ const version = await prisma.documentVersion.findUnique({
+ where: { id: documentVersionId },
+ select: { fileId: true },
+ });
+
+ let fileId = version?.fileId;
+
+ // Process document if no fileId exists
+ if (!fileId) {
+ metadata.set("status", "processing");
+ metadata.set("step", "Processing document...");
+ metadata.set("progress", 10);
+
+ // Route to appropriate processing task based on content type
+ if (PDF_CONTENT_TYPES.includes(contentType)) {
+ const result = await processPdfForAITask.triggerAndWait({
+ documentId,
+ documentVersionId,
+ teamId,
+ documentName,
+ filePath,
+ storageType,
+ contentType,
+ });
+
+ if (!result.ok) {
+ metadata.set("status", "failed");
+ metadata.set("step", "PDF processing failed");
+ metadata.set("error", "Failed to process PDF");
+ throw new Error("Failed to process PDF");
+ }
+
+ fileId = result.output.fileId;
+ } else if (EXCEL_CONTENT_TYPES.includes(contentType)) {
+ const result = await processExcelForAITask.triggerAndWait({
+ documentId,
+ documentVersionId,
+ teamId,
+ documentName,
+ filePath,
+ storageType,
+ contentType,
+ });
+
+ if (!result.ok) {
+ metadata.set("status", "failed");
+ metadata.set("step", "Excel processing failed");
+ metadata.set("error", "Failed to process Excel file");
+ throw new Error("Failed to process Excel file");
+ }
+
+ fileId = result.output.fileId;
+ } else if (IMAGE_CONTENT_TYPES.includes(contentType)) {
+ const result = await processImageForAITask.triggerAndWait({
+ documentId,
+ documentVersionId,
+ teamId,
+ documentName,
+ filePath,
+ storageType,
+ contentType,
+ });
+
+ if (!result.ok) {
+ metadata.set("status", "failed");
+ metadata.set("step", "Image processing failed");
+ metadata.set("error", "Failed to process image");
+ throw new Error("Failed to process image");
+ }
+
+ fileId = result.output.fileId;
+ }
+ } else {
+ logger.info("Document already processed, reusing fileId", {
+ documentId,
+ fileId,
+ });
+ metadata.set("step", "Using existing processed file...");
+ metadata.set("progress", 50);
+ }
+
+ if (!fileId) {
+ metadata.set("status", "failed");
+ metadata.set("step", "Processing failed");
+ metadata.set("error", "No file ID returned from processing");
+ throw new Error("No file ID returned from processing");
+ }
+
+ metadata.set("fileId", fileId);
+
+ // Add file to vector store
+ metadata.set("status", "indexing");
+ metadata.set("step", "Adding to vector store...");
+ metadata.set("progress", 80);
+
+ const vsResult = await addFileToVectorStoreTask.triggerAndWait({
+ fileId,
+ vectorStoreId,
+ metadata: fileMetadata,
+ });
+
+ if (!vsResult.ok) {
+ metadata.set("status", "failed");
+ metadata.set("step", "Failed to add to vector store");
+ metadata.set("error", "Failed to add file to vector store");
+ throw new Error("Failed to add file to vector store");
+ }
+
+ const { vectorStoreFileId } = vsResult.output;
+
+ // Update document version with vectorStoreFileId
+ await prisma.documentVersion.update({
+ where: { id: documentVersionId },
+ data: { vectorStoreFileId },
+ });
+
+ metadata.set("status", "completed");
+ metadata.set("step", "Indexed successfully");
+ metadata.set("progress", 100);
+ metadata.set("vectorStoreFileId", vectorStoreFileId);
+
+ logger.info("Document processed and indexed successfully", {
+ documentId,
+ fileId,
+ vectorStoreFileId,
+ });
+
+ return { vectorStoreFileId, fileId };
+ },
+});
diff --git a/ee/features/ai/lib/trigger/process-excel-for-ai.ts b/ee/features/ai/lib/trigger/process-excel-for-ai.ts
new file mode 100644
index 00000000..7a51e3b8
--- /dev/null
+++ b/ee/features/ai/lib/trigger/process-excel-for-ai.ts
@@ -0,0 +1,193 @@
+import { openai } from "@/ee/features/ai/lib/models/openai";
+import { logger, metadata, task } from "@trigger.dev/sdk/v3";
+import path from "path";
+import * as XLSX from "xlsx";
+
+import { getFile } from "@/lib/files/get-file";
+import { putFileServer } from "@/lib/files/put-file-server";
+import prisma from "@/lib/prisma";
+
+import type { ProcessFilePayload } from "./types";
+
+/**
+ * Convert Excel workbook to Markdown format
+ */
+function excelToMarkdown(workbook: XLSX.WorkBook): string {
+ const markdown: string[] = [];
+
+ for (const sheetName of workbook.SheetNames) {
+ const sheet = workbook.Sheets[sheetName];
+ if (!sheet) continue;
+
+ // Add sheet header
+ markdown.push(`## ${sheetName}\n`);
+
+ // Convert sheet to array of arrays
+ const data: unknown[][] = XLSX.utils.sheet_to_json(sheet, {
+ header: 1,
+ defval: "",
+ });
+
+ if (data.length === 0) {
+ markdown.push("*Empty sheet*\n");
+ continue;
+ }
+
+ // Get headers (first row)
+ const headers = data[0] as string[];
+ if (!headers || headers.length === 0) continue;
+
+ // Create markdown table header
+ markdown.push(
+ "| " + headers.map((h) => String(h || "")).join(" | ") + " |",
+ );
+ markdown.push("| " + headers.map(() => "---").join(" | ") + " |");
+
+ // Add data rows
+ for (let i = 1; i < data.length; i++) {
+ const row = data[i] as unknown[];
+ if (!row) continue;
+
+ const cells = headers.map((_, idx) => {
+ const cell = row[idx];
+ // Handle different cell types
+ if (cell === null || cell === undefined) return "";
+ if (typeof cell === "object") return JSON.stringify(cell);
+ return String(cell);
+ });
+
+ markdown.push("| " + cells.join(" | ") + " |");
+ }
+
+ markdown.push(""); // Empty line between sheets
+ }
+
+ return markdown.join("\n");
+}
+
+/**
+ * Process an Excel file for AI indexing
+ * Converts Excel to Markdown, saves to S3, then uploads to OpenAI
+ */
+export const processExcelForAITask = task({
+ id: "process-excel-for-ai",
+ retry: { maxAttempts: 3 },
+ queue: {
+ concurrencyLimit: 5,
+ },
+ run: async (
+ payload: ProcessFilePayload,
+ ): Promise<{ fileId: string; markdownPath?: string }> => {
+ const {
+ documentId,
+ documentVersionId,
+ teamId,
+ documentName,
+ filePath,
+ storageType,
+ } = payload;
+
+ logger.info("Processing Excel for AI", {
+ documentId,
+ documentVersionId,
+ teamId,
+ });
+
+ // Check if fileId already exists
+ const version = await prisma.documentVersion.findUnique({
+ where: { id: documentVersionId },
+ select: { fileId: true },
+ });
+
+ if (version?.fileId) {
+ logger.info("Excel already processed, reusing fileId", {
+ documentId,
+ fileId: version.fileId,
+ });
+ return { fileId: version.fileId };
+ }
+
+ metadata.set("status", "retrieving");
+ metadata.set("step", "Retrieving Excel file...");
+ metadata.set("progress", 10);
+
+ // Get file URL
+ const fileUrl = await getFile({
+ type: storageType,
+ data: filePath,
+ isDownload: true,
+ });
+
+ // Fetch file
+ const response = await fetch(fileUrl);
+ if (!response.ok) {
+ throw new Error(`Failed to fetch Excel file: ${response.statusText}`);
+ }
+
+ const buffer = Buffer.from(await response.arrayBuffer());
+
+ metadata.set("status", "processing");
+ metadata.set("step", "Converting Excel to Markdown...");
+ metadata.set("progress", 30);
+
+ // Parse Excel file
+ const workbook = XLSX.read(buffer, { type: "buffer" });
+
+ // Convert to markdown
+ const markdown = excelToMarkdown(workbook);
+
+ // Add document metadata to markdown
+ const fullMarkdown = `# ${documentName}\n\n${markdown}`;
+
+ metadata.set("step", "Saving processed content...");
+ metadata.set("progress", 50);
+
+ // Save markdown to S3 alongside original file
+ // Derive markdown path from original file path
+ const markdownPath = filePath.replace(/\.[^.]+$/, ".ai.md");
+ const markdownBuffer = Buffer.from(fullMarkdown, "utf-8");
+
+ // Extract docId from file path
+ const match = filePath.match(/(doc_[^\/]+)\//);
+ const docId = match ? match[1] : undefined;
+
+ // Save markdown file to S3
+ await putFileServer({
+ file: {
+ name: path.basename(markdownPath),
+ type: "text/markdown",
+ buffer: markdownBuffer,
+ },
+ teamId,
+ docId,
+ restricted: false,
+ });
+
+ metadata.set("status", "uploading");
+ metadata.set("step", "Uploading to OpenAI...");
+ metadata.set("progress", 70);
+
+ // Upload markdown to OpenAI Files
+ const file = new File([markdownBuffer], `${documentName}.md`, {
+ type: "text/markdown",
+ });
+ const fileResponse = await openai.files.create({
+ file,
+ purpose: "assistants",
+ });
+
+ // Update document version with fileId
+ await prisma.documentVersion.update({
+ where: { id: documentVersionId },
+ data: { fileId: fileResponse.id },
+ });
+
+ logger.info("Excel processed successfully", {
+ documentId,
+ fileId: fileResponse.id,
+ markdownPath,
+ });
+
+ return { fileId: fileResponse.id, markdownPath };
+ },
+});
diff --git a/ee/features/ai/lib/trigger/process-image-for-ai.ts b/ee/features/ai/lib/trigger/process-image-for-ai.ts
new file mode 100644
index 00000000..eb1161ac
--- /dev/null
+++ b/ee/features/ai/lib/trigger/process-image-for-ai.ts
@@ -0,0 +1,173 @@
+import { vertex } from "@/ee/features/ai/lib/models/google";
+import { openai } from "@/ee/features/ai/lib/models/openai";
+import { logger, metadata, task } from "@trigger.dev/sdk/v3";
+import { generateText } from "ai";
+import path from "path";
+
+import { getFile } from "@/lib/files/get-file";
+import { putFileServer } from "@/lib/files/put-file-server";
+import prisma from "@/lib/prisma";
+
+import type { ProcessFilePayload } from "./types";
+
+const IMAGE_ANALYSIS_PROMPT = `Analyze this image thoroughly and provide a detailed description for document search purposes.
+
+Include the following in your analysis:
+1. **Type of Image**: What kind of image is this? (e.g., diagram, chart, photo, screenshot, infographic, flowchart, etc.)
+2. **Main Content**: What is the primary subject or content of the image?
+3. **Text Content**: Extract and transcribe any visible text, labels, titles, or captions.
+4. **Data & Numbers**: If there are charts, graphs, or tables, describe the data being presented.
+5. **Key Details**: Note any important visual elements, colors, symbols, or indicators that convey meaning.
+6. **Context**: What is the likely purpose or context of this image? What document or presentation might it belong to?
+
+Format your response as clear, searchable text that would help someone find this image when searching for related topics.`;
+
+/**
+ * Process an image file for AI indexing
+ * Uses Google Vertex AI (Gemini Flash) to analyze the image and generate a description
+ */
+export const processImageForAITask = task({
+ id: "process-image-for-ai",
+ retry: { maxAttempts: 3 },
+ queue: {
+ concurrencyLimit: 5,
+ },
+ run: async (
+ payload: ProcessFilePayload,
+ ): Promise<{ fileId: string; markdownPath?: string }> => {
+ const {
+ documentId,
+ documentVersionId,
+ teamId,
+ documentName,
+ filePath,
+ storageType,
+ contentType,
+ } = payload;
+
+ logger.info("Processing image for AI", {
+ documentId,
+ documentVersionId,
+ teamId,
+ contentType,
+ });
+
+ // Check if fileId already exists
+ const version = await prisma.documentVersion.findUnique({
+ where: { id: documentVersionId },
+ select: { fileId: true },
+ });
+
+ if (version?.fileId) {
+ logger.info("Image already processed, reusing fileId", {
+ documentId,
+ fileId: version.fileId,
+ });
+ return { fileId: version.fileId };
+ }
+
+ metadata.set("status", "retrieving");
+ metadata.set("step", "Retrieving image...");
+ metadata.set("progress", 10);
+
+ // Get file URL
+ const fileUrl = await getFile({
+ type: storageType,
+ data: filePath,
+ isDownload: true,
+ });
+
+ // Fetch file
+ const response = await fetch(fileUrl);
+ if (!response.ok) {
+ throw new Error(`Failed to fetch image: ${response.statusText}`);
+ }
+
+ const imageBuffer = Buffer.from(await response.arrayBuffer());
+
+ metadata.set("status", "processing");
+ metadata.set("step", "Analyzing image with AI...");
+ metadata.set("progress", 30);
+
+ // Use Gemini Flash to analyze the image
+ const { text: imageDescription } = await generateText({
+ model: vertex("gemini-3-flash-preview"),
+ messages: [
+ {
+ role: "user",
+ content: [
+ {
+ type: "image",
+ image: imageBuffer,
+ },
+ {
+ type: "text",
+ text: IMAGE_ANALYSIS_PROMPT,
+ },
+ ],
+ },
+ ],
+ });
+
+ metadata.set("step", "Saving processed content...");
+ metadata.set("progress", 50);
+
+ // Create markdown content with image analysis
+ const fullMarkdown = `# ${documentName}
+
+## Image Analysis
+
+${imageDescription}
+
+---
+*This content was automatically generated from image analysis.*
+`;
+
+ // Save markdown to S3 alongside original file
+ const markdownPath = filePath.replace(/\.[^.]+$/, ".ai.md");
+ const markdownBuffer = Buffer.from(fullMarkdown, "utf-8");
+
+ // Extract docId from file path
+ const match = filePath.match(/(doc_[^\/]+)\//);
+ const docId = match ? match[1] : undefined;
+
+ // Save markdown file to S3
+ await putFileServer({
+ file: {
+ name: path.basename(markdownPath),
+ type: "text/markdown",
+ buffer: markdownBuffer,
+ },
+ teamId,
+ docId,
+ restricted: false,
+ });
+
+ metadata.set("status", "uploading");
+ metadata.set("step", "Uploading to OpenAI...");
+ metadata.set("progress", 70);
+
+ // Upload markdown to OpenAI Files
+ const file = new File([markdownBuffer], `${documentName}.md`, {
+ type: "text/markdown",
+ });
+ const fileResponse = await openai.files.create({
+ file,
+ purpose: "assistants",
+ });
+
+ // Update document version with fileId
+ await prisma.documentVersion.update({
+ where: { id: documentVersionId },
+ data: { fileId: fileResponse.id },
+ });
+
+ logger.info("Image processed successfully", {
+ documentId,
+ fileId: fileResponse.id,
+ markdownPath,
+ });
+
+ return { fileId: fileResponse.id, markdownPath };
+ },
+});
diff --git a/ee/features/ai/lib/trigger/process-pdf-for-ai.ts b/ee/features/ai/lib/trigger/process-pdf-for-ai.ts
new file mode 100644
index 00000000..b883b30b
--- /dev/null
+++ b/ee/features/ai/lib/trigger/process-pdf-for-ai.ts
@@ -0,0 +1,88 @@
+import { logger, metadata, task } from "@trigger.dev/sdk/v3";
+import path from "path";
+
+import { openai } from "@/ee/features/ai/lib/models/openai";
+import { getFile } from "@/lib/files/get-file";
+import prisma from "@/lib/prisma";
+
+import type { ProcessFilePayload } from "./types";
+
+/**
+ * Process a PDF file for AI indexing
+ * PDFs can be uploaded directly to OpenAI without conversion
+ */
+export const processPdfForAITask = task({
+ id: "process-pdf-for-ai",
+ retry: { maxAttempts: 3 },
+ queue: {
+ concurrencyLimit: 5,
+ },
+ run: async (payload: ProcessFilePayload): Promise<{ fileId: string }> => {
+ const { documentId, documentVersionId, teamId, documentName, filePath, storageType } =
+ payload;
+
+ logger.info("Processing PDF for AI", {
+ documentId,
+ documentVersionId,
+ teamId,
+ });
+
+ // Check if fileId already exists
+ const version = await prisma.documentVersion.findUnique({
+ where: { id: documentVersionId },
+ select: { fileId: true },
+ });
+
+ if (version?.fileId) {
+ logger.info("PDF already processed, reusing fileId", {
+ documentId,
+ fileId: version.fileId,
+ });
+ return { fileId: version.fileId };
+ }
+
+ metadata.set("status", "retrieving");
+ metadata.set("step", "Retrieving PDF...");
+ metadata.set("progress", 20);
+
+ // Get file URL
+ const fileUrl = await getFile({
+ type: storageType,
+ data: filePath,
+ isDownload: true,
+ });
+
+ // Fetch file
+ const response = await fetch(fileUrl);
+ if (!response.ok) {
+ throw new Error(`Failed to fetch PDF: ${response.statusText}`);
+ }
+
+ const buffer = Buffer.from(await response.arrayBuffer());
+ const fileName = path.basename(filePath);
+
+ metadata.set("status", "uploading");
+ metadata.set("step", "Uploading to OpenAI...");
+ metadata.set("progress", 60);
+
+ // Upload to OpenAI Files
+ const file = new File([buffer], fileName, { type: "application/pdf" });
+ const fileResponse = await openai.files.create({
+ file,
+ purpose: "assistants",
+ });
+
+ // Update document version with fileId
+ await prisma.documentVersion.update({
+ where: { id: documentVersionId },
+ data: { fileId: fileResponse.id },
+ });
+
+ logger.info("PDF processed successfully", {
+ documentId,
+ fileId: fileResponse.id,
+ });
+
+ return { fileId: fileResponse.id };
+ },
+});
diff --git a/ee/features/ai/lib/trigger/types.ts b/ee/features/ai/lib/trigger/types.ts
new file mode 100644
index 00000000..43832d5b
--- /dev/null
+++ b/ee/features/ai/lib/trigger/types.ts
@@ -0,0 +1,84 @@
+import { DocumentStorageType } from "@prisma/client";
+
+export type AIProcessingStatus =
+ | "pending"
+ | "retrieving"
+ | "processing"
+ | "uploading"
+ | "indexing"
+ | "completed"
+ | "failed";
+
+export type AIProcessingMetadata = {
+ status: AIProcessingStatus;
+ documentName: string;
+ documentId: string;
+ step: string;
+ progress: number;
+ error?: string;
+ vectorStoreFileId?: string;
+ fileId?: string;
+};
+
+export type ProcessDocumentPayload = {
+ documentId: string;
+ documentVersionId: string;
+ teamId: string;
+ vectorStoreId: string;
+ documentName: string;
+ filePath: string;
+ storageType: DocumentStorageType;
+ contentType: string;
+ metadata: {
+ teamId: string;
+ documentId: string;
+ documentName: string;
+ versionId: string;
+ folderId?: string;
+ dataroomId?: string;
+ dataroomDocumentId?: string;
+ dataroomFolderId?: string;
+ };
+};
+
+export type ProcessFilePayload = {
+ documentId: string;
+ documentVersionId: string;
+ teamId: string;
+ documentName: string;
+ filePath: string;
+ storageType: DocumentStorageType;
+ contentType: string;
+};
+
+export type AddToVectorStorePayload = {
+ fileId: string;
+ vectorStoreId: string;
+ metadata: {
+ teamId: string;
+ documentId: string;
+ documentName: string;
+ versionId: string;
+ folderId?: string;
+ dataroomId?: string;
+ dataroomDocumentId?: string;
+ dataroomFolderId?: string;
+ };
+};
+
+// Supported content types
+export const PDF_CONTENT_TYPES = ["application/pdf"];
+
+export const EXCEL_CONTENT_TYPES = [
+ "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+ "application/vnd.ms-excel",
+ "text/csv",
+];
+
+export const IMAGE_CONTENT_TYPES = ["image/jpeg", "image/png", "image/webp"];
+
+export const SUPPORTED_AI_CONTENT_TYPES = [
+ ...PDF_CONTENT_TYPES,
+ ...EXCEL_CONTENT_TYPES,
+ ...IMAGE_CONTENT_TYPES,
+];
diff --git a/lib/constants.ts b/lib/constants.ts
index cbcbf264..f5b5f849 100644
--- a/lib/constants.ts
+++ b/lib/constants.ts
@@ -83,6 +83,7 @@ export const SUPPORTED_DOCUMENT_MIME_TYPES = [
"application/rtf", // .rtf
"text/rtf", // .rtf
"text/plain", // .txt
+ "text/markdown", // .md
"image/vnd.dwg", // .dwg
"image/vnd.dxf", // .dxf
"image/png", // .png
diff --git a/package-lock.json b/package-lock.json
index b865b911..3466b131 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -9,6 +9,7 @@
"version": "0.1.0",
"hasInstallScript": true,
"dependencies": {
+ "@ai-sdk/google-vertex": "^4.0.1",
"@ai-sdk/openai": "^3.0.1",
"@ai-sdk/react": "^3.0.3",
"@aws-sdk/client-lambda": "^3.958.0",
@@ -177,6 +178,22 @@
"node": ">=22"
}
},
+ "node_modules/@ai-sdk/anthropic": {
+ "version": "3.0.1",
+ "resolved": "https://registry.npmjs.org/@ai-sdk/anthropic/-/anthropic-3.0.1.tgz",
+ "integrity": "sha512-MOiwKs76ilEmau/WRMnGWlheTUoB+cbvXCse+SAtpW5ATLreInsuYlspLABn12Dxu3w1Xzke1dT+tmEnxhy9SA==",
+ "license": "Apache-2.0",
+ "dependencies": {
+ "@ai-sdk/provider": "3.0.0",
+ "@ai-sdk/provider-utils": "4.0.1"
+ },
+ "engines": {
+ "node": ">=18"
+ },
+ "peerDependencies": {
+ "zod": "^3.25.76 || ^4.1.8"
+ }
+ },
"node_modules/@ai-sdk/gateway": {
"version": "3.0.2",
"resolved": "https://registry.npmjs.org/@ai-sdk/gateway/-/gateway-3.0.2.tgz",
@@ -194,6 +211,41 @@
"zod": "^3.25.76 || ^4.1.8"
}
},
+ "node_modules/@ai-sdk/google": {
+ "version": "3.0.1",
+ "resolved": "https://registry.npmjs.org/@ai-sdk/google/-/google-3.0.1.tgz",
+ "integrity": "sha512-gh7i4lEvd1CElmefkq7+RoUhNkhP2OTshzVxSt7/Vh2AV5wTPLhduKJMg1c7SFwErytqffO3el/M/LlfCsqzEw==",
+ "license": "Apache-2.0",
+ "dependencies": {
+ "@ai-sdk/provider": "3.0.0",
+ "@ai-sdk/provider-utils": "4.0.1"
+ },
+ "engines": {
+ "node": ">=18"
+ },
+ "peerDependencies": {
+ "zod": "^3.25.76 || ^4.1.8"
+ }
+ },
+ "node_modules/@ai-sdk/google-vertex": {
+ "version": "4.0.1",
+ "resolved": "https://registry.npmjs.org/@ai-sdk/google-vertex/-/google-vertex-4.0.1.tgz",
+ "integrity": "sha512-WVyHy8cLlFHng3nmC3d1a/eRWZJ2zQjQkKss1kMZnU5x7M5FtAS97xKjf5UtKWvswaI9+AfcM632Dc1Hmu3M0w==",
+ "license": "Apache-2.0",
+ "dependencies": {
+ "@ai-sdk/anthropic": "3.0.1",
+ "@ai-sdk/google": "3.0.1",
+ "@ai-sdk/provider": "3.0.0",
+ "@ai-sdk/provider-utils": "4.0.1",
+ "google-auth-library": "^10.5.0"
+ },
+ "engines": {
+ "node": ">=18"
+ },
+ "peerDependencies": {
+ "zod": "^3.25.76 || ^4.1.8"
+ }
+ },
"node_modules/@ai-sdk/openai": {
"version": "3.0.1",
"resolved": "https://registry.npmjs.org/@ai-sdk/openai/-/openai-3.0.1.tgz",
@@ -9522,6 +9574,15 @@
"node": ">=0.6"
}
},
+ "node_modules/bignumber.js": {
+ "version": "9.3.1",
+ "resolved": "https://registry.npmjs.org/bignumber.js/-/bignumber.js-9.3.1.tgz",
+ "integrity": "sha512-Ko0uX15oIUS7wJ3Rb30Fs6SkVbLmPBAKdlm7q9+ak9bbIeFf0MwuBsQV6z7+X768/cHsfg+WlysDWJcmthjsjQ==",
+ "license": "MIT",
+ "engines": {
+ "node": "*"
+ }
+ },
"node_modules/binary": {
"version": "0.3.0",
"resolved": "https://registry.npmjs.org/binary/-/binary-0.3.0.tgz",
@@ -10916,6 +10977,15 @@
"resolved": "https://registry.npmjs.org/damerau-levenshtein/-/damerau-levenshtein-1.0.8.tgz",
"integrity": "sha512-sdQSFB7+llfUcQHUQO3+B8ERRj0Oa4w9POWMI/puGtuf7gFywGmkaLCElnudfTiKZV+NvHqL0ifzdrI8Ro7ESA=="
},
+ "node_modules/data-uri-to-buffer": {
+ "version": "4.0.1",
+ "resolved": "https://registry.npmjs.org/data-uri-to-buffer/-/data-uri-to-buffer-4.0.1.tgz",
+ "integrity": "sha512-0R9ikRb668HB7QDxT1vkpuUBtqc53YyAwMwGeUFKRojY/NWKvdZ+9UYtRfGmhqNbRkTSVpMbmyhXipFFv2cb/A==",
+ "license": "MIT",
+ "engines": {
+ "node": ">= 12"
+ }
+ },
"node_modules/data-view-buffer": {
"version": "1.0.1",
"resolved": "https://registry.npmjs.org/data-view-buffer/-/data-view-buffer-1.0.1.tgz",
@@ -12513,6 +12583,29 @@
"reusify": "^1.0.4"
}
},
+ "node_modules/fetch-blob": {
+ "version": "3.2.0",
+ "resolved": "https://registry.npmjs.org/fetch-blob/-/fetch-blob-3.2.0.tgz",
+ "integrity": "sha512-7yAQpD2UMJzLi1Dqv7qFYnPbaPx7ZfFK6PiIxQ4PfkGPyNyl2Ugx+a/umUonmKqjhM4DnfbMvdX6otXq83soQQ==",
+ "funding": [
+ {
+ "type": "github",
+ "url": "https://github.com/sponsors/jimmywarting"
+ },
+ {
+ "type": "paypal",
+ "url": "https://paypal.me/jimmywarting"
+ }
+ ],
+ "license": "MIT",
+ "dependencies": {
+ "node-domexception": "^1.0.0",
+ "web-streams-polyfill": "^3.0.3"
+ },
+ "engines": {
+ "node": "^12.20 || >= 14.13"
+ }
+ },
"node_modules/fflate": {
"version": "0.4.8",
"resolved": "https://registry.npmjs.org/fflate/-/fflate-0.4.8.tgz",
@@ -12640,6 +12733,18 @@
"url": "https://github.com/sponsors/isaacs"
}
},
+ "node_modules/formdata-polyfill": {
+ "version": "4.0.10",
+ "resolved": "https://registry.npmjs.org/formdata-polyfill/-/formdata-polyfill-4.0.10.tgz",
+ "integrity": "sha512-buewHzMvYL29jdeQTVILecSaZKnt/RJWjoZCF5OW60Z67/GmSLBkOFM7qh1PI3zFNtJbaZL5eQu1vLfazOwj4g==",
+ "license": "MIT",
+ "dependencies": {
+ "fetch-blob": "^3.1.2"
+ },
+ "engines": {
+ "node": ">=12.20.0"
+ }
+ },
"node_modules/fraction.js": {
"version": "5.3.4",
"resolved": "https://registry.npmjs.org/fraction.js/-/fraction.js-5.3.4.tgz",
@@ -12880,6 +12985,90 @@
"node": ">=8"
}
},
+ "node_modules/gaxios": {
+ "version": "7.1.3",
+ "resolved": "https://registry.npmjs.org/gaxios/-/gaxios-7.1.3.tgz",
+ "integrity": "sha512-YGGyuEdVIjqxkxVH1pUTMY/XtmmsApXrCVv5EU25iX6inEPbV+VakJfLealkBtJN69AQmh1eGOdCl9Sm1UP6XQ==",
+ "license": "Apache-2.0",
+ "dependencies": {
+ "extend": "^3.0.2",
+ "https-proxy-agent": "^7.0.1",
+ "node-fetch": "^3.3.2",
+ "rimraf": "^5.0.1"
+ },
+ "engines": {
+ "node": ">=18"
+ }
+ },
+ "node_modules/gaxios/node_modules/agent-base": {
+ "version": "7.1.4",
+ "resolved": "https://registry.npmjs.org/agent-base/-/agent-base-7.1.4.tgz",
+ "integrity": "sha512-MnA+YT8fwfJPgBx3m60MNqakm30XOkyIoH1y6huTQvC0PwZG7ki8NacLBcrPbNoo8vEZy7Jpuk7+jMO+CUovTQ==",
+ "license": "MIT",
+ "engines": {
+ "node": ">= 14"
+ }
+ },
+ "node_modules/gaxios/node_modules/https-proxy-agent": {
+ "version": "7.0.6",
+ "resolved": "https://registry.npmjs.org/https-proxy-agent/-/https-proxy-agent-7.0.6.tgz",
+ "integrity": "sha512-vK9P5/iUfdl95AI+JVyUuIcVtd4ofvtrOr3HNtM2yxC9bnMbEdp3x01OhQNnjb8IJYi38VlTE3mBXwcfvywuSw==",
+ "license": "MIT",
+ "dependencies": {
+ "agent-base": "^7.1.2",
+ "debug": "4"
+ },
+ "engines": {
+ "node": ">= 14"
+ }
+ },
+ "node_modules/gaxios/node_modules/node-fetch": {
+ "version": "3.3.2",
+ "resolved": "https://registry.npmjs.org/node-fetch/-/node-fetch-3.3.2.tgz",
+ "integrity": "sha512-dRB78srN/l6gqWulah9SrxeYnxeddIG30+GOqK/9OlLVyLg3HPnr6SqOWTWOXKRwC2eGYCkZ59NNuSgvSrpgOA==",
+ "license": "MIT",
+ "dependencies": {
+ "data-uri-to-buffer": "^4.0.0",
+ "fetch-blob": "^3.1.4",
+ "formdata-polyfill": "^4.0.10"
+ },
+ "engines": {
+ "node": "^12.20.0 || ^14.13.1 || >=16.0.0"
+ },
+ "funding": {
+ "type": "opencollective",
+ "url": "https://opencollective.com/node-fetch"
+ }
+ },
+ "node_modules/gaxios/node_modules/rimraf": {
+ "version": "5.0.10",
+ "resolved": "https://registry.npmjs.org/rimraf/-/rimraf-5.0.10.tgz",
+ "integrity": "sha512-l0OE8wL34P4nJH/H2ffoaniAokM2qSmrtXHmlpvYr5AVVX8msAyW0l8NVJFDxlSK4u3Uh/f41cQheDVdnYijwQ==",
+ "license": "ISC",
+ "dependencies": {
+ "glob": "^10.3.7"
+ },
+ "bin": {
+ "rimraf": "dist/esm/bin.mjs"
+ },
+ "funding": {
+ "url": "https://github.com/sponsors/isaacs"
+ }
+ },
+ "node_modules/gcp-metadata": {
+ "version": "8.1.2",
+ "resolved": "https://registry.npmjs.org/gcp-metadata/-/gcp-metadata-8.1.2.tgz",
+ "integrity": "sha512-zV/5HKTfCeKWnxG0Dmrw51hEWFGfcF2xiXqcA3+J90WDuP0SvoiSO5ORvcBsifmx/FoIjgQN3oNOGaQ5PhLFkg==",
+ "license": "Apache-2.0",
+ "dependencies": {
+ "gaxios": "^7.0.0",
+ "google-logging-utils": "^1.0.0",
+ "json-bigint": "^1.0.0"
+ },
+ "engines": {
+ "node": ">=18"
+ }
+ },
"node_modules/generic-pool": {
"version": "3.9.0",
"resolved": "https://registry.npmjs.org/generic-pool/-/generic-pool-3.9.0.tgz",
@@ -13060,6 +13249,33 @@
"url": "https://github.com/sponsors/ljharb"
}
},
+ "node_modules/google-auth-library": {
+ "version": "10.5.0",
+ "resolved": "https://registry.npmjs.org/google-auth-library/-/google-auth-library-10.5.0.tgz",
+ "integrity": "sha512-7ABviyMOlX5hIVD60YOfHw4/CxOfBhyduaYB+wbFWCWoni4N7SLcV46hrVRktuBbZjFC9ONyqamZITN7q3n32w==",
+ "license": "Apache-2.0",
+ "dependencies": {
+ "base64-js": "^1.3.0",
+ "ecdsa-sig-formatter": "^1.0.11",
+ "gaxios": "^7.0.0",
+ "gcp-metadata": "^8.0.0",
+ "google-logging-utils": "^1.0.0",
+ "gtoken": "^8.0.0",
+ "jws": "^4.0.0"
+ },
+ "engines": {
+ "node": ">=18"
+ }
+ },
+ "node_modules/google-logging-utils": {
+ "version": "1.1.3",
+ "resolved": "https://registry.npmjs.org/google-logging-utils/-/google-logging-utils-1.1.3.tgz",
+ "integrity": "sha512-eAmLkjDjAFCVXg7A1unxHsLf961m6y17QFqXqAXGj/gVkKFrEICfStRfwUlGNfeCEjNRa32JEWOUTlYXPyyKvA==",
+ "license": "Apache-2.0",
+ "engines": {
+ "node": ">=14"
+ }
+ },
"node_modules/gopd": {
"version": "1.2.0",
"resolved": "https://registry.npmjs.org/gopd/-/gopd-1.2.0.tgz",
@@ -13082,6 +13298,19 @@
"resolved": "https://registry.npmjs.org/graphemer/-/graphemer-1.4.0.tgz",
"integrity": "sha512-EtKwoO6kxCL9WO5xipiHTZlSzBm7WLT627TqC/uVRd0HKmq8NXyebnNYxDoBi7wt8eTWrUrKXCOVaFq9x1kgag=="
},
+ "node_modules/gtoken": {
+ "version": "8.0.0",
+ "resolved": "https://registry.npmjs.org/gtoken/-/gtoken-8.0.0.tgz",
+ "integrity": "sha512-+CqsMbHPiSTdtSO14O51eMNlrp9N79gmeqmXeouJOhfucAedHw9noVe/n5uJk3tbKE6a+6ZCQg3RPhVhHByAIw==",
+ "license": "MIT",
+ "dependencies": {
+ "gaxios": "^7.0.0",
+ "jws": "^4.0.0"
+ },
+ "engines": {
+ "node": ">=18"
+ }
+ },
"node_modules/hachure-fill": {
"version": "0.5.2",
"resolved": "https://registry.npmjs.org/hachure-fill/-/hachure-fill-0.5.2.tgz",
@@ -14333,6 +14562,15 @@
"node": ">=6"
}
},
+ "node_modules/json-bigint": {
+ "version": "1.0.0",
+ "resolved": "https://registry.npmjs.org/json-bigint/-/json-bigint-1.0.0.tgz",
+ "integrity": "sha512-SiPv/8VpZuWbvLSMtTDU8hEfrZWg/mH/nV/b4o0CYbSxu1UIQPLdwKOCIyLQX+VIPO5vrLX3i8qtqFyhdPSUSQ==",
+ "license": "MIT",
+ "dependencies": {
+ "bignumber.js": "^9.0.0"
+ }
+ },
"node_modules/json-buffer": {
"version": "3.0.1",
"resolved": "https://registry.npmjs.org/json-buffer/-/json-buffer-3.0.1.tgz",
@@ -16580,6 +16818,26 @@
"node": "^10 || ^12 || >=14"
}
},
+ "node_modules/node-domexception": {
+ "version": "1.0.0",
+ "resolved": "https://registry.npmjs.org/node-domexception/-/node-domexception-1.0.0.tgz",
+ "integrity": "sha512-/jKZoMpw0F8GRwl4/eLROPA3cfcXtLApP0QzLmUT/HuPCZWyB7IY9ZrMeKw2O/nFIqPQB3PVM9aYm0F312AXDQ==",
+ "deprecated": "Use your platform's native DOMException instead",
+ "funding": [
+ {
+ "type": "github",
+ "url": "https://github.com/sponsors/jimmywarting"
+ },
+ {
+ "type": "github",
+ "url": "https://paypal.me/jimmywarting"
+ }
+ ],
+ "license": "MIT",
+ "engines": {
+ "node": ">=10.5.0"
+ }
+ },
"node_modules/node-fetch": {
"version": "2.7.0",
"resolved": "https://registry.npmjs.org/node-fetch/-/node-fetch-2.7.0.tgz",
@@ -21730,6 +21988,15 @@
"url": "https://github.com/sponsors/wooorm"
}
},
+ "node_modules/web-streams-polyfill": {
+ "version": "3.3.3",
+ "resolved": "https://registry.npmjs.org/web-streams-polyfill/-/web-streams-polyfill-3.3.3.tgz",
+ "integrity": "sha512-d2JWLCivmZYTSIoge9MsgFCZrt571BikcWGYkjC1khllbTeDlGqZ2D8vD8E/lJa8WGWbb7Plm8/XJYV7IJHZZw==",
+ "license": "MIT",
+ "engines": {
+ "node": ">= 8"
+ }
+ },
"node_modules/web-vitals": {
"version": "4.2.4",
"resolved": "https://registry.npmjs.org/web-vitals/-/web-vitals-4.2.4.tgz",
diff --git a/package.json b/package.json
index 9deb63c4..73cbfb69 100644
--- a/package.json
+++ b/package.json
@@ -20,6 +20,7 @@
"dev:prisma": "npx prisma generate && npx prisma migrate deploy"
},
"dependencies": {
+ "@ai-sdk/google-vertex": "^4.0.1",
"@ai-sdk/openai": "^3.0.1",
"@ai-sdk/react": "^3.0.3",
"@aws-sdk/client-lambda": "^3.958.0",
diff --git a/pages/api/teams/[teamId]/datarooms/[id]/documents/index.ts b/pages/api/teams/[teamId]/datarooms/[id]/documents/index.ts
index 70f13952..5e8513e5 100644
--- a/pages/api/teams/[teamId]/datarooms/[id]/documents/index.ts
+++ b/pages/api/teams/[teamId]/datarooms/[id]/documents/index.ts
@@ -1,11 +1,17 @@
import { NextApiRequest, NextApiResponse } from "next";
import { isTeamPausedById } from "@/ee/features/billing/cancellation/lib/is-team-paused";
+import {
+ addFileToVectorStoreTask,
+ processDocumentForAITask,
+ SUPPORTED_AI_CONTENT_TYPES,
+} from "@/ee/features/ai/lib/trigger";
import { authOptions } from "@/pages/api/auth/[...nextauth]";
import { runs } from "@trigger.dev/sdk/v3";
import { waitUntil } from "@vercel/functions";
import { getServerSession } from "next-auth/next";
+import { getFeatureFlags } from "@/lib/featureFlags";
import { errorhandler } from "@/lib/errorHandler";
import prisma from "@/lib/prisma";
import { sendDataroomChangeNotificationTask } from "@/lib/trigger/dataroom-change-notification";
@@ -164,16 +170,28 @@ export default async function handle(
},
});
- const document = await prisma.dataroomDocument.create({
+ const dataroomDocument = await prisma.dataroomDocument.create({
data: {
documentId,
dataroomId,
folderId: folder?.id,
},
include: {
+ document: {
+ include: {
+ versions: {
+ where: { isPrimary: true },
+ take: 1,
+ },
+ },
+ },
dataroom: {
select: {
+ teamId: true,
+ name: true,
enableChangeNotifications: true,
+ agentsEnabled: true,
+ vectorStoreId: true,
links: {
select: { id: true },
orderBy: { createdAt: "desc" },
@@ -187,8 +205,79 @@ export default async function handle(
},
});
+ // Auto-index document if dataroom has AI agents enabled
+ if (
+ dataroomDocument.dataroom.agentsEnabled &&
+ dataroomDocument.dataroom.vectorStoreId
+ ) {
+ const primaryVersion = dataroomDocument.document.versions[0];
+ const contentType = primaryVersion?.contentType || "";
+
+ // Check if AI feature is enabled for the team
+ const features = await getFeatureFlags({ teamId });
+
+ if (features.ai && primaryVersion && SUPPORTED_AI_CONTENT_TYPES.includes(contentType)) {
+ const filePath =
+ primaryVersion.originalFile && contentType !== "application/pdf"
+ ? primaryVersion.originalFile
+ : primaryVersion.file;
+
+ const fileMetadata = {
+ teamId: dataroomDocument.dataroom.teamId,
+ documentId: dataroomDocument.document.id,
+ documentName: dataroomDocument.document.name,
+ versionId: primaryVersion.id,
+ dataroomId: dataroomDocument.dataroomId,
+ dataroomDocumentId: dataroomDocument.id,
+ dataroomFolderId: dataroomDocument.folderId || "root",
+ };
+
+ try {
+ // If document already has fileId, just add to vector store
+ if (primaryVersion.fileId) {
+ waitUntil(
+ addFileToVectorStoreTask.trigger({
+ fileId: primaryVersion.fileId,
+ vectorStoreId: dataroomDocument.dataroom.vectorStoreId,
+ metadata: fileMetadata,
+ }),
+ );
+ } else {
+ // Trigger full processing
+ waitUntil(
+ processDocumentForAITask.trigger(
+ {
+ documentId: dataroomDocument.document.id,
+ documentVersionId: primaryVersion.id,
+ teamId: dataroomDocument.dataroom.teamId,
+ vectorStoreId: dataroomDocument.dataroom.vectorStoreId,
+ documentName: dataroomDocument.document.name,
+ filePath,
+ storageType: primaryVersion.storageType,
+ contentType,
+ metadata: fileMetadata,
+ },
+ {
+ idempotencyKey: `ai-index-dataroom-${dataroomId}-${primaryVersion.id}`,
+ tags: [
+ `team_${teamId}`,
+ `dataroom_${dataroomId}`,
+ `document_${dataroomDocument.document.id}`,
+ `version_${primaryVersion.id}`,
+ ],
+ },
+ ),
+ );
+ }
+ } catch (error) {
+ console.error("Error triggering AI indexing for document:", error);
+ // Don't fail the document add, just log the error
+ }
+ }
+ }
+
// Check if the team has the dataroom change notification enabled
- if (document.dataroom.enableChangeNotifications) {
+ if (dataroomDocument.dataroom.enableChangeNotifications) {
// Get all delayed and queued runs for this dataroom
const allRuns = await runs.list({
taskIdentifier: ["send-dataroom-change-notification"],
@@ -204,16 +293,16 @@ export default async function handle(
sendDataroomChangeNotificationTask.trigger(
{
dataroomId,
- dataroomDocumentId: document.id,
+ dataroomDocumentId: dataroomDocument.id,
senderUserId: userId,
teamId,
},
{
- idempotencyKey: `dataroom-notification-${teamId}-${dataroomId}-${document.id}`,
+ idempotencyKey: `dataroom-notification-${teamId}-${dataroomId}-${dataroomDocument.id}`,
tags: [
`team_${teamId}`,
`dataroom_${dataroomId}`,
- `document_${document.id}`,
+ `document_${dataroomDocument.id}`,
],
delay: new Date(Date.now() + 10 * 60 * 1000), // 10 minute delay
},
@@ -221,7 +310,7 @@ export default async function handle(
);
}
- return res.status(201).json(document);
+ return res.status(201).json(dataroomDocument);
} catch (error) {
log({
message: `Failed to create dataroom document. \n\n*teamId*: _${teamId}_, \n\n*dataroomId*: ${dataroomId} \n\n ${error}`,
diff --git a/pages/api/teams/[teamId]/documents/[id]/add-to-dataroom.ts b/pages/api/teams/[teamId]/documents/[id]/add-to-dataroom.ts
index 18225405..3c01b187 100644
--- a/pages/api/teams/[teamId]/documents/[id]/add-to-dataroom.ts
+++ b/pages/api/teams/[teamId]/documents/[id]/add-to-dataroom.ts
@@ -1,12 +1,24 @@
import { NextApiRequest, NextApiResponse } from "next";
+import {
+ addFileToVectorStoreTask,
+ processDocumentForAITask,
+ SUPPORTED_AI_CONTENT_TYPES,
+} from "@/ee/features/ai/lib/trigger";
import { authOptions } from "@/pages/api/auth/[...nextauth]";
+import { waitUntil } from "@vercel/functions";
import { getServerSession } from "next-auth/next";
+import { getFeatureFlags } from "@/lib/featureFlags";
import { errorhandler } from "@/lib/errorHandler";
import prisma from "@/lib/prisma";
import { CustomUser } from "@/lib/types";
+export const config = {
+ // in order to enable `waitUntil` function
+ supportsResponseStreaming: true,
+};
+
export default async function handle(
req: NextApiRequest,
res: NextApiResponse,
@@ -59,17 +71,47 @@ export default async function handle(
});
}
- try {
- await prisma.dataroom.update({
- where: {
- id: dataroomId,
+ // Fetch dataroom with AI settings
+ const dataroom = await prisma.dataroom.findUnique({
+ where: { id: dataroomId },
+ select: {
+ id: true,
+ teamId: true,
+ name: true,
+ agentsEnabled: true,
+ vectorStoreId: true,
+ },
+ });
+
+ if (!dataroom) {
+ return res.status(404).json({
+ message: "Dataroom not found!",
+ });
+ }
+
+ // Fetch document with primary version
+ const document = await prisma.document.findUnique({
+ where: { id: docId },
+ include: {
+ versions: {
+ where: { isPrimary: true },
+ take: 1,
},
+ },
+ });
+
+ if (!document) {
+ return res.status(404).json({
+ message: "Document not found!",
+ });
+ }
+
+ let dataroomDocument;
+ try {
+ dataroomDocument = await prisma.dataroomDocument.create({
data: {
- documents: {
- create: {
- documentId: docId,
- },
- },
+ documentId: docId,
+ dataroomId,
},
});
} catch (error) {
@@ -78,6 +120,74 @@ export default async function handle(
});
}
+ // Auto-index document if dataroom has AI agents enabled
+ if (dataroom.agentsEnabled && dataroom.vectorStoreId) {
+ const primaryVersion = document.versions[0];
+ const contentType = primaryVersion?.contentType || "";
+
+ // Check if AI feature is enabled for the team
+ const features = await getFeatureFlags({ teamId });
+
+ if (features.ai && primaryVersion && SUPPORTED_AI_CONTENT_TYPES.includes(contentType)) {
+ const filePath =
+ primaryVersion.originalFile && contentType !== "application/pdf"
+ ? primaryVersion.originalFile
+ : primaryVersion.file;
+
+ const fileMetadata = {
+ teamId: dataroom.teamId,
+ documentId: document.id,
+ documentName: document.name,
+ versionId: primaryVersion.id,
+ dataroomId: dataroom.id,
+ dataroomDocumentId: dataroomDocument.id,
+ dataroomFolderId: "root",
+ };
+
+ try {
+ // If document already has fileId, just add to vector store
+ if (primaryVersion.fileId) {
+ waitUntil(
+ addFileToVectorStoreTask.trigger({
+ fileId: primaryVersion.fileId,
+ vectorStoreId: dataroom.vectorStoreId,
+ metadata: fileMetadata,
+ }),
+ );
+ } else {
+ // Trigger full processing
+ waitUntil(
+ processDocumentForAITask.trigger(
+ {
+ documentId: document.id,
+ documentVersionId: primaryVersion.id,
+ teamId: dataroom.teamId,
+ vectorStoreId: dataroom.vectorStoreId,
+ documentName: document.name,
+ filePath,
+ storageType: primaryVersion.storageType,
+ contentType,
+ metadata: fileMetadata,
+ },
+ {
+ idempotencyKey: `ai-index-dataroom-${dataroomId}-${primaryVersion.id}`,
+ tags: [
+ `team_${teamId}`,
+ `dataroom_${dataroomId}`,
+ `document_${document.id}`,
+ `version_${primaryVersion.id}`,
+ ],
+ },
+ ),
+ );
+ }
+ } catch (error) {
+ console.error("Error triggering AI indexing for document:", error);
+ // Don't fail the document add, just log the error
+ }
+ }
+ }
+
return res.status(200).json({
message: "Document added to dataroom!",
});
diff --git a/trigger.config.ts b/trigger.config.ts
index bc170916..4c0cc7fa 100644
--- a/trigger.config.ts
+++ b/trigger.config.ts
@@ -4,7 +4,7 @@ import { defineConfig, timeout } from "@trigger.dev/sdk/v3";
export default defineConfig({
project: "proj_plmsfqvqunboixacjjus",
- dirs: ["./lib/trigger"],
+ dirs: ["./lib/trigger", "./ee/features/ai/lib/trigger"],
maxDuration: timeout.None, // no max duration
retries: {
enabledInDev: false,