📎 feat: Upload as Text Support for Plaintext, STT, RAG, and Token Limits (#8868)

* 🪶 feat: Add Support for Uploading Plaintext Files feat: delineate between OCR and text handling in fileConfig field of config file - also adds support for passing in mimetypes as just plain file extensions feat: add showLabel bool to support future synthetic component DynamicDropdownInput feat: add new combination dropdown-input component in params panel to support file type token limits refactor: move hovercard to side to align with other hovercards chore: clean up autogenerated comments feat: add delineation to file upload path between text and ocr configured filetypes feat: add token limit checks during file upload refactor: move textParsing out of ocrEnabled logic refactor: clean up types for filetype config refactor: finish decoupling DynamicDropdownInput from fileTokenLimits fix: move image token cost function into file to fix circular dependency causing unittest to fail and remove unused var for linter chore: remove out of scope code following review refactor: make fileTokenLimit conform to existing styles chore: remove unused localization string chore: undo changes to DynamicInput and other strays feat: add fileTokenLimit to all provider config panels fix: move textParsing back into ocr tool_resource block for now so that it doesn't interfere with other upload types * 📤 feat: Add RAG API Endpoint Support for Text Parsing (#8849) * feat: implement RAG API integration for text parsing with fallback to native parsing * chore: remove TODO now that placeholder and fllback are implemented * ✈️ refactor: Migrate Text Parsing to TS (#8892) * refactor: move generateShortLivedToken to packages/api * refactor: move textParsing logic into packages/api * refactor: reduce nesting and dry code with createTextFile * fix: add proper source handling * fix: mock new parseText and parseTextNative functions in jest file * ci: add test coverage for textParser * 💬 feat: Add Audio File Support to Upload as Text (#8893) * feat: add STT support for Upload as Text * refactor: move processAudioFile to packages/api * refactor: move textParsing from utils to files * fix: remove audio/mp3 from unsupported mimetypes test since it is now supported * ✂️ feat: Configurable File Token Limits and Truncation (#8911) * feat: add configurable fileTokenLimit default value * fix: add stt to fileConfig merge logic * fix: add fileTokenLimit to mergeFileConfig logic so configurable value is actually respected from yaml * feat: add token limiting to parsed text files * fix: add extraction logic and update tests so fileTokenLimit isnt sent to LLM providers * fix: address comments * refactor: rename textTokenLimiter.ts to text.ts * chore: update form-data package to address CVE-2025-7783 and update package-lock * feat: use default supported mime types for ocr on frontend file validation * fix: should be using logger.debug not console.debug * fix: mock existsSync in text.spec.ts * fix: mock logger rather than every one of its function calls * fix: reorganize imports and streamline file upload processing logic * refactor: update createTextFile function to use destructured parameters and improve readability * chore: update file validation to use EToolResources for improved type safety * chore: update import path for types in audio processing module * fix: update file configuration access and replace console.debug with logger.debug for improved logging --------- Co-authored-by: Dustin Healy <dustinhealy1@gmail.com> Co-authored-by: Dustin Healy <54083382+dustinhealy@users.noreply.github.com>
2025-09-22 06:00:56 +02:00 · 2025-08-27 03:44:39 -04:00 · 2025-08-27 03:44:39 -04:00 · 48f6f8f2f8
commit 48f6f8f2f8
parent 74bc0440f0
41 changed files with 847 additions and 75 deletions
--- a/api/app/clients/prompts/createContextHandlers.js
+++ b/api/app/clients/prompts/createContextHandlers.js
@ -1,7 +1,6 @@
 const axios = require('axios');
-const { isEnabled } = require('@librechat/api');
 const { logger } = require('@librechat/data-schemas');
-const { generateShortLivedToken } = require('~/server/services/AuthService');
+const { isEnabled, generateShortLivedToken } = require('@librechat/api');

 const footer = `Use the context as your learned knowledge to better answer the user.

--- a/api/app/clients/tools/util/fileSearch.js
+++ b/api/app/clients/tools/util/fileSearch.js
@ -2,9 +2,9 @@ const { z } = require('zod');
 const axios = require('axios');
 const { tool } = require('@langchain/core/tools');
 const { logger } = require('@librechat/data-schemas');
+const { generateShortLivedToken } = require('@librechat/api');
 const { Tools, EToolResources } = require('librechat-data-provider');
 const { filterFilesByAgentAccess } = require('~/server/services/Files/permissions');
-const { generateShortLivedToken } = require('~/server/services/AuthService');
 const { getFiles } = require('~/models/File');

 /**
--- a/api/server/routes/files/files.js
+++ b/api/server/routes/files/files.js
@ -387,7 +387,8 @@ router.post('/', async (req, res) => {

    if (
      error.message?.includes('Invalid file format') ||
-      error.message?.includes('No OCR result')
+      error.message?.includes('No OCR result') ||
+      error.message?.includes('exceeds token limit')
    ) {
      message = error.message;
    }
--- a/api/server/routes/files/images.js
+++ b/api/server/routes/files/images.js
@ -29,6 +29,17 @@ router.post('/', async (req, res) => {
  } catch (error) {
    // TODO: delete remote file if it exists
    logger.error('[/files/images] Error processing file:', error);
+
+    let message = 'Error processing file';
+
+    if (
+      error.message?.includes('Invalid file format') ||
+      error.message?.includes('No OCR result') ||
+      error.message?.includes('exceeds token limit')
+    ) {
+      message = error.message;
+    }
+
    try {
      const filepath = path.join(
        appConfig.paths.imageOutput,
@ -39,7 +50,7 @@ router.post('/', async (req, res) => {
    } catch (error) {
      logger.error('[/files/images] Error deleting file:', error);
    }
-    res.status(500).json({ message: 'Error processing file' });
+    res.status(500).json({ message });
  } finally {
    try {
      await fs.unlink(req.file.path);
--- a/api/server/services/AuthService.js
+++ b/api/server/services/AuthService.js
@ -500,18 +500,6 @@ const resendVerificationEmail = async (req) => {
    };
  }
 };
-/**
- * Generate a short-lived JWT token
- * @param {String} userId - The ID of the user
- * @param {String} [expireIn='5m'] - The expiration time for the token (default is 5 minutes)
- * @returns {String} - The generated JWT token
- */
-const generateShortLivedToken = (userId, expireIn = '5m') => {
-  return jwt.sign({ id: userId }, process.env.JWT_SECRET, {
-    expiresIn: expireIn,
-    algorithm: 'HS256',
-  });
-};

 module.exports = {
  logoutUser,
@ -522,5 +510,4 @@ module.exports = {
  setOpenIDAuthTokens,
  requestPasswordReset,
  resendVerificationEmail,
-  generateShortLivedToken,
 };
--- a/api/server/services/Endpoints/anthropic/build.js
+++ b/api/server/services/Endpoints/anthropic/build.js
@ -6,6 +6,7 @@ const buildOptions = (endpoint, parsedBody) => {
    modelLabel,
    promptPrefix,
    maxContextTokens,
+    fileTokenLimit,
    resendFiles = anthropicSettings.resendFiles.default,
    promptCache = anthropicSettings.promptCache.default,
    thinking = anthropicSettings.thinking.default,
@ -29,6 +30,7 @@ const buildOptions = (endpoint, parsedBody) => {
    greeting,
    spec,
    maxContextTokens,
+    fileTokenLimit,
    modelOptions,
  });

--- a/api/server/services/Endpoints/bedrock/build.js
+++ b/api/server/services/Endpoints/bedrock/build.js
@ -6,6 +6,7 @@ const buildOptions = (endpoint, parsedBody) => {
    modelLabel: name,
    promptPrefix,
    maxContextTokens,
+    fileTokenLimit,
    resendFiles = true,
    imageDetail,
    iconURL,
@ -24,6 +25,7 @@ const buildOptions = (endpoint, parsedBody) => {
    spec,
    promptPrefix,
    maxContextTokens,
+    fileTokenLimit,
    model_parameters,
  });

--- a/api/server/services/Endpoints/custom/build.js
+++ b/api/server/services/Endpoints/custom/build.js
@ -7,6 +7,7 @@ const buildOptions = (endpoint, parsedBody, endpointType) => {
    chatGptLabel,
    promptPrefix,
    maxContextTokens,
+    fileTokenLimit,
    resendFiles = true,
    imageDetail,
    iconURL,
@ -27,6 +28,7 @@ const buildOptions = (endpoint, parsedBody, endpointType) => {
    greeting,
    spec,
    maxContextTokens,
+    fileTokenLimit,
    modelOptions,
  });

--- a/api/server/services/Endpoints/google/build.js
+++ b/api/server/services/Endpoints/google/build.js
@ -12,6 +12,7 @@ const buildOptions = (endpoint, parsedBody) => {
    spec,
    artifacts,
    maxContextTokens,
+    fileTokenLimit,
    ...modelOptions
  } = parsedBody;
  const endpointOption = removeNullishValues({
@ -24,6 +25,7 @@ const buildOptions = (endpoint, parsedBody) => {
    greeting,
    spec,
    maxContextTokens,
+    fileTokenLimit,
    modelOptions,
  });

--- a/api/server/services/Endpoints/openAI/build.js
+++ b/api/server/services/Endpoints/openAI/build.js
@ -7,6 +7,7 @@ const buildOptions = (endpoint, parsedBody) => {
    chatGptLabel,
    promptPrefix,
    maxContextTokens,
+    fileTokenLimit,
    resendFiles = true,
    imageDetail,
    iconURL,
@ -27,6 +28,7 @@ const buildOptions = (endpoint, parsedBody) => {
    greeting,
    spec,
    maxContextTokens,
+    fileTokenLimit,
    modelOptions,
  });

--- a/api/server/services/Files/Audio/STTService.js
+++ b/api/server/services/Files/Audio/STTService.js
@ -319,4 +319,4 @@ async function speechToText(req, res) {
  await sttService.processSpeechToText(req, res);
 }

-module.exports = { speechToText };
+module.exports = { STTService, speechToText };
--- a/api/server/services/Files/Local/crud.js
+++ b/api/server/services/Files/Local/crud.js
@ -3,7 +3,7 @@ const path = require('path');
 const axios = require('axios');
 const { logger } = require('@librechat/data-schemas');
 const { EModelEndpoint } = require('librechat-data-provider');
-const { generateShortLivedToken } = require('~/server/services/AuthService');
+const { generateShortLivedToken } = require('@librechat/api');
 const { getBufferMetadata } = require('~/server/utils');
 const paths = require('~/config/paths');

--- a/api/server/services/Files/VectorDB/crud.js
+++ b/api/server/services/Files/VectorDB/crud.js
@ -1,10 +1,9 @@
 const fs = require('fs');
 const axios = require('axios');
 const FormData = require('form-data');
-const { logAxiosError } = require('@librechat/api');
 const { logger } = require('@librechat/data-schemas');
 const { FileSources } = require('librechat-data-provider');
-const { generateShortLivedToken } = require('~/server/services/AuthService');
+const { logAxiosError, generateShortLivedToken } = require('@librechat/api');

 /**
 * Deletes a file from the vector database. This function takes a file object, constructs the full path, and
--- a/api/server/services/Files/images/encode.js
+++ b/api/server/services/Files/images/encode.js
@ -1,13 +1,16 @@
 const axios = require('axios');
-const { logAxiosError } = require('@librechat/api');
+const { logger } = require('@librechat/data-schemas');
+const { logAxiosError, processTextWithTokenLimit } = require('@librechat/api');
 const {
  FileSources,
  VisionModes,
  ImageDetail,
  ContentTypes,
  EModelEndpoint,
+  mergeFileConfig,
 } = require('librechat-data-provider');
 const { getStrategyFunctions } = require('~/server/services/Files/strategies');
+const countTokens = require('~/server/utils/countTokens');

 /**
 * Converts a readable stream to a base64 encoded string.
@ -102,11 +105,28 @@ async function encodeAndFormat(req, files, endpoint, mode) {
    return result;
  }

+  const fileTokenLimit =
+    req.body?.fileTokenLimit ?? mergeFileConfig(req.config?.fileConfig).fileTokenLimit;
+
  for (let file of files) {
    /** @type {FileSources} */
    const source = file.source ?? FileSources.local;
    if (source === FileSources.text && file.text) {
-      result.text += `${!result.text ? 'Attached document(s):\n```md' : '\n\n---\n\n'}# "${file.filename}"\n${file.text}\n`;
+      let fileText = file.text;
+
+      const { text: limitedText, wasTruncated } = await processTextWithTokenLimit({
+        text: fileText,
+        tokenLimit: fileTokenLimit,
+        tokenCountFn: (text) => countTokens(text),
+      });
+
+      if (wasTruncated) {
+        logger.debug(
+          `[encodeAndFormat] Text content truncated for file: ${file.filename} due to token limits`,
+        );
+      }
+
+      result.text += `${!result.text ? 'Attached document(s):\n```md' : '\n\n---\n\n'}# "${file.filename}"\n${limitedText}\n`;
    }

    if (!file.height) {
@ -135,7 +155,7 @@ async function encodeAndFormat(req, files, endpoint, mode) {
        base64Data = null;
        continue;
      } catch (error) {
-        // Error handling code
+        logger.error('Error processing image from blob storage:', error);
      }
    } else if (source !== FileSources.local && base64Only.has(endpoint)) {
      const [_file, imageURL] = await preparePayload(req, file);
--- a/api/server/services/Files/process.js
+++ b/api/server/services/Files/process.js
@ -17,7 +17,8 @@ const {
  isAssistantsEndpoint,
 } = require('librechat-data-provider');
 const { EnvVar } = require('@librechat/agents');
-const { sanitizeFilename } = require('@librechat/api');
+const { logger } = require('@librechat/data-schemas');
+const { sanitizeFilename, parseText, processAudioFile } = require('@librechat/api');
 const {
  convertImage,
  resizeAndConvert,
@ -33,7 +34,7 @@ const { checkCapability } = require('~/server/services/Config');
 const { LB_QueueAsyncCall } = require('~/server/utils/queue');
 const { getStrategyFunctions } = require('./strategies');
 const { determineFileType } = require('~/server/utils');
-const { logger } = require('~/config');
+const { STTService } = require('./Audio/STTService');

 /**
 * Creates a modular file upload wrapper that ensures filename sanitization
@ -552,35 +553,27 @@ const processAgentFileUpload = async ({ req, res, metadata }) => {
    }
    // Note: File search processing continues to dual storage logic below
  } else if (tool_resource === EToolResources.ocr) {
-    const isOCREnabled = await checkCapability(req, AgentCapabilities.ocr);
-    if (!isOCREnabled) {
-      throw new Error('OCR capability is not enabled for Agents');
-    }
-
-    const { handleFileUpload: uploadOCR } = getStrategyFunctions(
-      appConfig?.ocr?.strategy ?? FileSources.mistral_ocr,
-    );
    const { file_id, temp_file_id = null } = metadata;

-    const {
-      text,
-      bytes,
-      // TODO: OCR images support?
-      images: _i,
-      filename,
-      filepath: ocrFileURL,
-    } = await uploadOCR({ req, appConfig, file, loadAuthValues });
-
+    /**
+     * @param {object} params
+     * @param {string} params.text
+     * @param {number} params.bytes
+     * @param {string} params.filepath
+     * @param {string} params.type
+     * @return {Promise<void>}
+     */
+    const createTextFile = async ({ text, bytes, filepath, type = 'text/plain' }) => {
      const fileInfo = removeNullishValues({
        text,
        bytes,
        file_id,
        temp_file_id,
        user: req.user.id,
-      type: 'text/plain',
-      filepath: ocrFileURL,
+        type,
+        filepath: filepath ?? file.path,
        source: FileSources.text,
-      filename: filename ?? file.originalname,
+        filename: file.originalname,
        model: messageAttachment ? undefined : req.body.model,
        context: messageAttachment ? FileContext.message_attachment : FileContext.agents,
      });
@ -597,6 +590,47 @@ const processAgentFileUpload = async ({ req, res, metadata }) => {
      return res
        .status(200)
        .json({ message: 'Agent file uploaded and processed successfully', ...result });
+    };
+
+    const fileConfig = mergeFileConfig(appConfig.fileConfig);
+
+    const shouldUseOCR = fileConfig.checkType(
+      file.mimetype,
+      fileConfig.ocr?.supportedMimeTypes || [],
+    );
+
+    if (shouldUseOCR && !(await checkCapability(req, AgentCapabilities.ocr))) {
+      throw new Error('OCR capability is not enabled for Agents');
+    } else if (shouldUseOCR) {
+      const { handleFileUpload: uploadOCR } = getStrategyFunctions(
+        appConfig?.ocr?.strategy ?? FileSources.mistral_ocr,
+      );
+      const { text, bytes, filepath: ocrFileURL } = await uploadOCR({ req, file, loadAuthValues });
+      return await createTextFile({ text, bytes, filepath: ocrFileURL });
+    }
+
+    const shouldUseSTT = fileConfig.checkType(
+      file.mimetype,
+      fileConfig.stt?.supportedMimeTypes || [],
+    );
+
+    if (shouldUseSTT) {
+      const sttService = await STTService.getInstance();
+      const { text, bytes } = await processAudioFile({ file, sttService });
+      return await createTextFile({ text, bytes });
+    }
+
+    const shouldUseText = fileConfig.checkType(
+      file.mimetype,
+      fileConfig.text?.supportedMimeTypes || [],
+    );
+
+    if (!shouldUseText) {
+      throw new Error(`File type ${file.mimetype} is not supported for OCR or text parsing`);
+    }
+
+    const { text, bytes } = await parseText({ req, file, file_id });
+    return await createTextFile({ text, bytes, type: file.mimetype });
  }

  // Dual storage pattern for RAG files: Storage + Vector DB
--- a/api/server/services/Files/processFiles.test.js
+++ b/api/server/services/Files/processFiles.test.js
@ -101,6 +101,11 @@ jest.mock('~/server/utils', () => ({
  determineFileType: jest.fn(),
 }));

+jest.mock('@librechat/api', () => ({
+  parseText: jest.fn(),
+  parseTextNative: jest.fn(),
+}));
+
 // Import the actual processFiles function after all mocks are set up
 const { processFiles } = require('./process');
 const { updateFileUsage } = require('~/models/File');
--- a/client/src/hooks/Files/useFileHandling.ts
+++ b/client/src/hooks/Files/useFileHandling.ts
@ -135,10 +135,14 @@ const useFileHandling = (params?: UseFileHandling) => {
        const file_id = body.get('file_id');
        clearUploadTimer(file_id as string);
        deleteFileById(file_id as string);
-        const errorMessage =
-          error?.code === 'ERR_CANCELED'
-            ? 'com_error_files_upload_canceled'
-            : (error?.response?.data?.message ?? 'com_error_files_upload');
+
+        let errorMessage = 'com_error_files_upload';
+
+        if (error?.code === 'ERR_CANCELED') {
+          errorMessage = 'com_error_files_upload_canceled';
+        } else if (error?.response?.data?.message) {
+          errorMessage = error.response.data.message;
+        }
        setError(errorMessage);
      },
    },
@ -256,6 +260,8 @@ const useFileHandling = (params?: UseFileHandling) => {
          fileConfig?.endpoints?.default ??
          defaultFileConfig.endpoints[endpoint] ??
          defaultFileConfig.endpoints.default,
+        toolResource: _toolResource,
+        fileConfig: fileConfig,
      });
    } catch (error) {
      console.error('file validation error', error);
--- a/client/src/locales/en/translation.json
+++ b/client/src/locales/en/translation.json
@ -889,6 +889,8 @@
  "com_ui_field_required": "This field is required",
  "com_ui_file_size": "File Size",
  "com_ui_files": "Files",
+  "com_ui_file_token_limit": "File Token Limit",
+  "com_ui_file_token_limit_desc": "Set maximum token limit for file processing to control costs and resource usage",
  "com_ui_filter_prompts": "Filter Prompts",
  "com_ui_filter_prompts_name": "Filter prompts by name",
  "com_ui_final_touch": "Final touch",
--- a/client/src/utils/files.ts
+++ b/client/src/utils/files.ts
@ -3,10 +3,11 @@ import {
  megabyte,
  QueryKeys,
  excelMimeTypes,
+  EToolResources,
  codeTypeMapping,
  fileConfig as defaultFileConfig,
 } from 'librechat-data-provider';
-import type { TFile, EndpointFileConfig } from 'librechat-data-provider';
+import type { TFile, EndpointFileConfig, FileConfig } from 'librechat-data-provider';
 import type { QueryClient } from '@tanstack/react-query';
 import type { ExtendedFile } from '~/common';

@ -203,11 +204,15 @@ export const validateFiles = ({
  fileList,
  setError,
  endpointFileConfig,
+  toolResource,
+  fileConfig,
 }: {
  fileList: File[];
  files: Map<string, ExtendedFile>;
  setError: (error: string) => void;
  endpointFileConfig: EndpointFileConfig;
+  toolResource?: string;
+  fileConfig: FileConfig | null;
 }) => {
  const { fileLimit, fileSizeLimit, totalSizeLimit, supportedMimeTypes } = endpointFileConfig;
  const existingFiles = Array.from(files.values());
@ -247,7 +252,16 @@ export const validateFiles = ({
      fileList[i] = newFile;
    }

-    if (!checkType(originalFile.type, supportedMimeTypes)) {
+    let mimeTypesToCheck = supportedMimeTypes;
+    if (toolResource === EToolResources.ocr) {
+      mimeTypesToCheck = [
+        ...(fileConfig?.text?.supportedMimeTypes || []),
+        ...(fileConfig?.ocr?.supportedMimeTypes || []),
+        ...(fileConfig?.stt?.supportedMimeTypes || []),
+      ];
+    }
+
+    if (!checkType(originalFile.type, mimeTypesToCheck)) {
      console.log(originalFile);
      setError('Currently, unsupported file type: ' + originalFile.type);
      return false;
--- a/package-lock.json
+++ b/package-lock.json
@ -51813,6 +51813,7 @@
        "@types/diff": "^6.0.0",
        "@types/express": "^5.0.0",
        "@types/jest": "^29.5.2",
+        "@types/jsonwebtoken": "^9.0.0",
        "@types/multer": "^1.4.13",
        "@types/node": "^20.3.0",
        "@types/react": "^18.2.18",
@ -51837,7 +51838,9 @@
        "diff": "^7.0.0",
        "eventsource": "^3.0.2",
        "express": "^4.21.2",
+        "form-data": "^4.0.4",
        "js-yaml": "^4.1.0",
+        "jsonwebtoken": "^9.0.0",
        "keyv": "^5.3.2",
        "librechat-data-provider": "*",
        "node-fetch": "2.7.0",
--- a/packages/api/package.json
+++ b/packages/api/package.json
@ -53,6 +53,7 @@
    "@types/diff": "^6.0.0",
    "@types/express": "^5.0.0",
    "@types/jest": "^29.5.2",
+    "@types/jsonwebtoken": "^9.0.0",
    "@types/multer": "^1.4.13",
    "@types/node": "^20.3.0",
    "@types/react": "^18.2.18",
@ -80,7 +81,9 @@
    "diff": "^7.0.0",
    "eventsource": "^3.0.2",
    "express": "^4.21.2",
+    "form-data": "^4.0.4",
    "js-yaml": "^4.1.0",
+    "jsonwebtoken": "^9.0.0",
    "keyv": "^5.3.2",
    "librechat-data-provider": "*",
    "node-fetch": "2.7.0",
--- a/packages/api/src/crypto/index.ts
+++ b/packages/api/src/crypto/index.ts
@ -1 +1,2 @@
 export * from './encryption';
+export * from './jwt';
--- a/packages/api/src/crypto/jwt.ts
+++ b/packages/api/src/crypto/jwt.ts
@ -0,0 +1,14 @@
+import jwt from 'jsonwebtoken';
+
+/**
+ * Generate a short-lived JWT token
+ * @param {String} userId - The ID of the user
+ * @param {String} [expireIn='5m'] - The expiration time for the token (default is 5 minutes)
+ * @returns {String} - The generated JWT token
+ */
+export const generateShortLivedToken = (userId: string, expireIn: string = '5m'): string => {
+  return jwt.sign({ id: userId }, process.env.JWT_SECRET!, {
+    expiresIn: expireIn,
+    algorithm: 'HS256',
+  });
+};
--- a/packages/api/src/files/audio.ts
+++ b/packages/api/src/files/audio.ts
@ -0,0 +1,38 @@
+import fs from 'fs';
+import { logger } from '@librechat/data-schemas';
+import type { STTService, AudioFileInfo, FileObject, AudioProcessingResult } from '~/types';
+
+/**
+ * Processes audio files using Speech-to-Text (STT) service.
+ * @param {Object} params - The parameters object.
+ * @param {FileObject} params.file - The audio file object.
+ * @param {STTService} params.sttService - The STT service instance.
+ * @returns {Promise<AudioProcessingResult>} A promise that resolves to an object containing text and bytes.
+ */
+export async function processAudioFile({
+  file,
+  sttService,
+}: {
+  file: FileObject;
+  sttService: STTService;
+}): Promise<AudioProcessingResult> {
+  try {
+    const audioBuffer = await fs.promises.readFile(file.path);
+    const audioFile: AudioFileInfo = {
+      originalname: file.originalname,
+      mimetype: file.mimetype,
+      size: file.size,
+    };
+
+    const [provider, sttSchema] = await sttService.getProviderSchema();
+    const text = await sttService.sttRequest(provider, sttSchema, { audioBuffer, audioFile });
+
+    return {
+      text,
+      bytes: Buffer.byteLength(text, 'utf8'),
+    };
+  } catch (error) {
+    logger.error('Error processing audio file with STT:', error);
+    throw new Error(`Failed to process audio file: ${(error as Error).message}`);
+  }
+}
--- a/packages/api/src/files/index.ts
+++ b/packages/api/src/files/index.ts
@ -1,2 +1,4 @@
 export * from './mistral/crud';
+export * from './audio';
+export * from './text';
 export * from './parse';
--- a/packages/api/src/files/text.spec.ts
+++ b/packages/api/src/files/text.spec.ts
@ -0,0 +1,255 @@
+import { FileSources } from 'librechat-data-provider';
+import { Readable } from 'stream';
+
+jest.mock('@librechat/data-schemas', () => ({
+  logger: {
+    debug: jest.fn(),
+    warn: jest.fn(),
+    error: jest.fn(),
+  },
+}));
+
+import { parseTextNative, parseText } from './text';
+
+jest.mock('fs', () => ({
+  readFileSync: jest.fn(),
+  createReadStream: jest.fn(),
+}));
+
+jest.mock('../crypto/jwt', () => ({
+  generateShortLivedToken: jest.fn(),
+}));
+
+jest.mock('axios', () => ({
+  get: jest.fn(),
+  post: jest.fn(),
+  interceptors: {
+    request: { use: jest.fn(), eject: jest.fn() },
+    response: { use: jest.fn(), eject: jest.fn() },
+  },
+}));
+
+jest.mock('form-data', () => {
+  return jest.fn().mockImplementation(() => ({
+    append: jest.fn(),
+    getHeaders: jest.fn().mockReturnValue({ 'content-type': 'multipart/form-data' }),
+  }));
+});
+
+import fs, { ReadStream } from 'fs';
+import axios from 'axios';
+import FormData from 'form-data';
+import { generateShortLivedToken } from '../crypto/jwt';
+
+const mockedFs = fs as jest.Mocked<typeof fs>;
+const mockedAxios = axios as jest.Mocked<typeof axios>;
+const mockedFormData = FormData as jest.MockedClass<typeof FormData>;
+const mockedGenerateShortLivedToken = generateShortLivedToken as jest.MockedFunction<
+  typeof generateShortLivedToken
+>;
+
+describe('text', () => {
+  const mockFile: Express.Multer.File = {
+    fieldname: 'file',
+    originalname: 'test.txt',
+    encoding: '7bit',
+    mimetype: 'text/plain',
+    size: 100,
+    destination: '/tmp',
+    filename: 'test.txt',
+    path: '/tmp/test.txt',
+    buffer: Buffer.from('test content'),
+    stream: new Readable(),
+  };
+
+  const mockReq = {
+    user: { id: 'user123' },
+  };
+
+  const mockFileId = 'file123';
+
+  beforeEach(() => {
+    jest.clearAllMocks();
+    delete process.env.RAG_API_URL;
+  });
+
+  describe('parseTextNative', () => {
+    it('should successfully parse a text file', () => {
+      const mockText = 'Hello, world!';
+      mockedFs.readFileSync.mockReturnValue(mockText);
+
+      const result = parseTextNative(mockFile);
+
+      expect(mockedFs.readFileSync).toHaveBeenCalledWith('/tmp/test.txt', 'utf8');
+      expect(result).toEqual({
+        text: mockText,
+        bytes: Buffer.byteLength(mockText, 'utf8'),
+        source: FileSources.text,
+      });
+    });
+
+    it('should throw an error when file cannot be read', () => {
+      const mockError = new Error('File not found');
+      mockedFs.readFileSync.mockImplementation(() => {
+        throw mockError;
+      });
+
+      expect(() => parseTextNative(mockFile)).toThrow(
+        'Failed to read file as text: Error: File not found',
+      );
+    });
+  });
+
+  describe('parseText', () => {
+    beforeEach(() => {
+      mockedGenerateShortLivedToken.mockReturnValue('mock-jwt-token');
+
+      const mockFormDataInstance = {
+        append: jest.fn(),
+        getHeaders: jest.fn().mockReturnValue({ 'content-type': 'multipart/form-data' }),
+      };
+      mockedFormData.mockImplementation(() => mockFormDataInstance as unknown as FormData);
+
+      mockedFs.createReadStream.mockReturnValue({} as unknown as ReadStream);
+    });
+
+    it('should fall back to native parsing when RAG_API_URL is not defined', async () => {
+      const mockText = 'Native parsing result';
+      mockedFs.readFileSync.mockReturnValue(mockText);
+
+      const result = await parseText({
+        req: mockReq,
+        file: mockFile,
+        file_id: mockFileId,
+      });
+
+      expect(result).toEqual({
+        text: mockText,
+        bytes: Buffer.byteLength(mockText, 'utf8'),
+        source: FileSources.text,
+      });
+      expect(mockedAxios.get).not.toHaveBeenCalled();
+    });
+
+    it('should fall back to native parsing when health check fails', async () => {
+      process.env.RAG_API_URL = 'http://rag-api.test';
+      const mockText = 'Native parsing result';
+      mockedFs.readFileSync.mockReturnValue(mockText);
+
+      mockedAxios.get.mockRejectedValue(new Error('Health check failed'));
+
+      const result = await parseText({
+        req: mockReq,
+        file: mockFile,
+        file_id: mockFileId,
+      });
+
+      expect(mockedAxios.get).toHaveBeenCalledWith('http://rag-api.test/health', {
+        timeout: 5000,
+      });
+      expect(result).toEqual({
+        text: mockText,
+        bytes: Buffer.byteLength(mockText, 'utf8'),
+        source: FileSources.text,
+      });
+    });
+
+    it('should fall back to native parsing when health check returns non-OK status', async () => {
+      process.env.RAG_API_URL = 'http://rag-api.test';
+      const mockText = 'Native parsing result';
+      mockedFs.readFileSync.mockReturnValue(mockText);
+
+      mockedAxios.get.mockResolvedValue({
+        status: 500,
+        statusText: 'Internal Server Error',
+      });
+
+      const result = await parseText({
+        req: mockReq,
+        file: mockFile,
+        file_id: mockFileId,
+      });
+
+      expect(result).toEqual({
+        text: mockText,
+        bytes: Buffer.byteLength(mockText, 'utf8'),
+        source: FileSources.text,
+      });
+    });
+
+    it('should accept empty text as valid RAG API response', async () => {
+      process.env.RAG_API_URL = 'http://rag-api.test';
+
+      mockedAxios.get.mockResolvedValue({
+        status: 200,
+        statusText: 'OK',
+      });
+
+      mockedAxios.post.mockResolvedValue({
+        data: {
+          text: '',
+        },
+      });
+
+      const result = await parseText({
+        req: mockReq,
+        file: mockFile,
+        file_id: mockFileId,
+      });
+
+      expect(result).toEqual({
+        text: '',
+        bytes: 0,
+        source: FileSources.text,
+      });
+    });
+
+    it('should fall back to native parsing when RAG API response lacks text property', async () => {
+      process.env.RAG_API_URL = 'http://rag-api.test';
+      const mockText = 'Native parsing result';
+      mockedFs.readFileSync.mockReturnValue(mockText);
+
+      mockedAxios.get.mockResolvedValue({
+        status: 200,
+        statusText: 'OK',
+      });
+
+      mockedAxios.post.mockResolvedValue({
+        data: {},
+      });
+
+      const result = await parseText({
+        req: mockReq,
+        file: mockFile,
+        file_id: mockFileId,
+      });
+
+      expect(result).toEqual({
+        text: mockText,
+        bytes: Buffer.byteLength(mockText, 'utf8'),
+        source: FileSources.text,
+      });
+    });
+
+    it('should fall back to native parsing when user is undefined', async () => {
+      process.env.RAG_API_URL = 'http://rag-api.test';
+      const mockText = 'Native parsing result';
+      mockedFs.readFileSync.mockReturnValue(mockText);
+
+      const result = await parseText({
+        req: { user: undefined },
+        file: mockFile,
+        file_id: mockFileId,
+      });
+
+      expect(mockedGenerateShortLivedToken).not.toHaveBeenCalled();
+      expect(mockedAxios.get).not.toHaveBeenCalled();
+      expect(mockedAxios.post).not.toHaveBeenCalled();
+      expect(result).toEqual({
+        text: mockText,
+        bytes: Buffer.byteLength(mockText, 'utf8'),
+        source: FileSources.text,
+      });
+    });
+  });
+});
--- a/packages/api/src/files/text.ts
+++ b/packages/api/src/files/text.ts
@ -0,0 +1,113 @@
+import fs from 'fs';
+import axios from 'axios';
+import FormData from 'form-data';
+import { logger } from '@librechat/data-schemas';
+import { FileSources } from 'librechat-data-provider';
+import type { Request as ServerRequest } from 'express';
+import { generateShortLivedToken } from '~/crypto/jwt';
+
+/**
+ * Attempts to parse text using RAG API, falls back to native text parsing
+ * @param {Object} params - The parameters object
+ * @param {Express.Request} params.req - The Express request object
+ * @param {Express.Multer.File} params.file - The uploaded file
+ * @param {string} params.file_id - The file ID
+ * @returns {Promise<{text: string, bytes: number, source: string}>}
+ */
+export async function parseText({
+  req,
+  file,
+  file_id,
+}: {
+  req: Pick<ServerRequest, 'user'> & {
+    user?: { id: string };
+  };
+  file: Express.Multer.File;
+  file_id: string;
+}): Promise<{ text: string; bytes: number; source: string }> {
+  if (!process.env.RAG_API_URL) {
+    logger.debug('[parseText] RAG_API_URL not defined, falling back to native text parsing');
+    return parseTextNative(file);
+  }
+
+  if (!req.user?.id) {
+    logger.debug('[parseText] No user ID provided, falling back to native text parsing');
+    return parseTextNative(file);
+  }
+
+  try {
+    const healthResponse = await axios.get(`${process.env.RAG_API_URL}/health`, {
+      timeout: 5000,
+    });
+    if (healthResponse?.statusText !== 'OK' && healthResponse?.status !== 200) {
+      logger.debug('[parseText] RAG API health check failed, falling back to native parsing');
+      return parseTextNative(file);
+    }
+  } catch (healthError) {
+    logger.debug(
+      '[parseText] RAG API health check failed, falling back to native parsing',
+      healthError,
+    );
+    return parseTextNative(file);
+  }
+
+  try {
+    const jwtToken = generateShortLivedToken(req.user.id);
+    const formData = new FormData();
+    formData.append('file_id', file_id);
+    formData.append('file', fs.createReadStream(file.path));
+
+    const formHeaders = formData.getHeaders();
+
+    const response = await axios.post(`${process.env.RAG_API_URL}/text`, formData, {
+      headers: {
+        Authorization: `Bearer ${jwtToken}`,
+        accept: 'application/json',
+        ...formHeaders,
+      },
+      timeout: 30000,
+    });
+
+    const responseData = response.data;
+    logger.debug('[parseText] Response from RAG API', responseData);
+
+    if (!('text' in responseData)) {
+      throw new Error('RAG API did not return parsed text');
+    }
+
+    return {
+      text: responseData.text,
+      bytes: Buffer.byteLength(responseData.text, 'utf8'),
+      source: FileSources.text,
+    };
+  } catch (error) {
+    logger.warn('[parseText] RAG API text parsing failed, falling back to native parsing', error);
+    return parseTextNative(file);
+  }
+}
+
+/**
+ * Native JavaScript text parsing fallback
+ * Simple text file reading - complex formats handled by RAG API
+ * @param {Express.Multer.File} file - The uploaded file
+ * @returns {{text: string, bytes: number, source: string}}
+ */
+export function parseTextNative(file: Express.Multer.File): {
+  text: string;
+  bytes: number;
+  source: string;
+} {
+  try {
+    const text = fs.readFileSync(file.path, 'utf8');
+    const bytes = Buffer.byteLength(text, 'utf8');
+
+    return {
+      text,
+      bytes,
+      source: FileSources.text,
+    };
+  } catch (error) {
+    console.error('[parseTextNative] Failed to parse file:', error);
+    throw new Error(`Failed to read file as text: ${error}`);
+  }
+}
--- a/packages/api/src/types/files.ts
+++ b/packages/api/src/types/files.ts
@ -0,0 +1,27 @@
+export interface STTService {
+  getInstance(): Promise<STTService>;
+  getProviderSchema(): Promise<[string, object]>;
+  sttRequest(
+    provider: string,
+    schema: object,
+    params: { audioBuffer: Buffer; audioFile: AudioFileInfo },
+  ): Promise<string>;
+}
+
+export interface AudioFileInfo {
+  originalname: string;
+  mimetype: string;
+  size: number;
+}
+
+export interface FileObject {
+  path: string;
+  originalname: string;
+  mimetype: string;
+  size: number;
+}
+
+export interface AudioProcessingResult {
+  text: string;
+  bytes: number;
+}
--- a/packages/api/src/types/index.ts
+++ b/packages/api/src/types/index.ts
@ -4,6 +4,7 @@ export * from './balance';
 export * from './endpoints';
 export * from './events';
 export * from './error';
+export * from './files';
 export * from './google';
 export * from './http';
 export * from './mistral';
--- a/packages/api/src/utils/index.ts
+++ b/packages/api/src/utils/index.ts
@ -11,6 +11,7 @@ export * from './llm';
 export * from './math';
 export * from './openid';
 export * from './tempChatRetention';
+export * from './text';
 export { default as Tokenizer } from './tokenizer';
 export * from './yaml';
 export * from './http';
--- a/packages/api/src/utils/llm.test.ts
+++ b/packages/api/src/utils/llm.test.ts
@ -7,6 +7,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(true);
    expect(result.promptPrefix).toBeUndefined();
    expect(result.maxContextTokens).toBeUndefined();
+    expect(result.fileTokenLimit).toBeUndefined();
    expect(result.modelLabel).toBeUndefined();
    expect(result.modelOptions).toEqual({});
  });
@ -17,6 +18,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(true);
    expect(result.promptPrefix).toBeUndefined();
    expect(result.maxContextTokens).toBeUndefined();
+    expect(result.fileTokenLimit).toBeUndefined();
    expect(result.modelLabel).toBeUndefined();
    expect(result.modelOptions).toEqual({});
  });
@ -26,6 +28,7 @@ describe('extractLibreChatParams', () => {
      resendFiles: false,
      promptPrefix: 'You are a helpful assistant',
      maxContextTokens: 4096,
+      fileTokenLimit: 50000,
      modelLabel: 'GPT-4',
      model: 'gpt-4',
      temperature: 0.7,
@ -37,6 +40,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(false);
    expect(result.promptPrefix).toBe('You are a helpful assistant');
    expect(result.maxContextTokens).toBe(4096);
+    expect(result.fileTokenLimit).toBe(50000);
    expect(result.modelLabel).toBe('GPT-4');
    expect(result.modelOptions).toEqual({
      model: 'gpt-4',
@ -50,6 +54,7 @@ describe('extractLibreChatParams', () => {
      resendFiles: true,
      promptPrefix: null,
      maxContextTokens: 2048,
+      fileTokenLimit: undefined,
      modelLabel: null,
      model: 'claude-3',
    };
@ -59,6 +64,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(true);
    expect(result.promptPrefix).toBeNull();
    expect(result.maxContextTokens).toBe(2048);
+    expect(result.fileTokenLimit).toBeUndefined();
    expect(result.modelLabel).toBeNull();
    expect(result.modelOptions).toEqual({
      model: 'claude-3',
@ -77,6 +83,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(true); // Should use default
    expect(result.promptPrefix).toBe('Test prefix');
    expect(result.maxContextTokens).toBeUndefined();
+    expect(result.fileTokenLimit).toBeUndefined();
    expect(result.modelLabel).toBeUndefined();
    expect(result.modelOptions).toEqual({
      model: 'gpt-3.5-turbo',
@ -90,6 +97,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(true); // Should use default
    expect(result.promptPrefix).toBeUndefined();
    expect(result.maxContextTokens).toBeUndefined();
+    expect(result.fileTokenLimit).toBeUndefined();
    expect(result.modelLabel).toBeUndefined();
    expect(result.modelOptions).toEqual({});
  });
@ -99,6 +107,7 @@ describe('extractLibreChatParams', () => {
      resendFiles: false,
      promptPrefix: 'Custom prompt',
      maxContextTokens: 8192,
+      fileTokenLimit: 25000,
      modelLabel: 'Custom Model',
      // Model options
      model: 'gpt-4',
@ -117,6 +126,7 @@ describe('extractLibreChatParams', () => {
    expect(result.resendFiles).toBe(false);
    expect(result.promptPrefix).toBe('Custom prompt');
    expect(result.maxContextTokens).toBe(8192);
+    expect(result.fileTokenLimit).toBe(25000);
    expect(result.modelLabel).toBe('Custom Model');

    // Model options should include everything else
--- a/packages/api/src/utils/llm.ts
+++ b/packages/api/src/utils/llm.ts
@ -8,6 +8,7 @@ type LibreChatParams = {
  resendFiles: boolean;
  promptPrefix?: string | null;
  maxContextTokens?: number;
+  fileTokenLimit?: number;
  modelLabel?: string | null;
 };

@ -32,6 +33,7 @@ export function extractLibreChatParams(
    (librechat.resendFiles.default as boolean);
  const promptPrefix = (delete modelOptions.promptPrefix, options.promptPrefix);
  const maxContextTokens = (delete modelOptions.maxContextTokens, options.maxContextTokens);
+  const fileTokenLimit = (delete modelOptions.fileTokenLimit, options.fileTokenLimit);
  const modelLabel = (delete modelOptions.modelLabel, options.modelLabel);

  return {
@ -40,6 +42,7 @@ export function extractLibreChatParams(
      LibreChatKeys
    >,
    maxContextTokens,
+    fileTokenLimit,
    promptPrefix,
    resendFiles,
    modelLabel,
--- a/packages/api/src/utils/text.ts
+++ b/packages/api/src/utils/text.ts
@ -0,0 +1,65 @@
+import { logger } from '@librechat/data-schemas';
+
+/**
+ * Processes text content by counting tokens and truncating if it exceeds the specified limit.
+ * @param text - The text content to process
+ * @param tokenLimit - The maximum number of tokens allowed
+ * @param tokenCountFn - Function to count tokens
+ * @returns Promise resolving to object with processed text, token count, and truncation status
+ */
+export async function processTextWithTokenLimit({
+  text,
+  tokenLimit,
+  tokenCountFn,
+}: {
+  text: string;
+  tokenLimit: number;
+  tokenCountFn: (text: string) => number;
+}): Promise<{ text: string; tokenCount: number; wasTruncated: boolean }> {
+  const originalTokenCount = await tokenCountFn(text);
+
+  if (originalTokenCount <= tokenLimit) {
+    return {
+      text,
+      tokenCount: originalTokenCount,
+      wasTruncated: false,
+    };
+  }
+
+  /**
+   * Doing binary search here to find the truncation point efficiently
+   * (May be a better way to go about this)
+   */
+  let low = 0;
+  let high = text.length;
+  let bestText = '';
+
+  logger.debug(
+    `[textTokenLimiter] Text content exceeds token limit: ${originalTokenCount} > ${tokenLimit}, truncating...`,
+  );
+
+  while (low <= high) {
+    const mid = Math.floor((low + high) / 2);
+    const truncatedText = text.substring(0, mid);
+    const tokenCount = await tokenCountFn(truncatedText);
+
+    if (tokenCount <= tokenLimit) {
+      bestText = truncatedText;
+      low = mid + 1;
+    } else {
+      high = mid - 1;
+    }
+  }
+
+  const finalTokenCount = await tokenCountFn(bestText);
+
+  logger.warn(
+    `[textTokenLimiter] Text truncated from ${originalTokenCount} to ${finalTokenCount} tokens (limit: ${tokenLimit})`,
+  );
+
+  return {
+    text: bestText,
+    tokenCount: finalTokenCount,
+    wasTruncated: true,
+  };
+}
--- a/packages/data-provider/specs/filetypes.spec.ts
+++ b/packages/data-provider/specs/filetypes.spec.ts
@ -14,7 +14,7 @@ import {
 } from '../src/file-config';

 describe('MIME Type Regex Patterns', () => {
-  const unsupportedMimeTypes = ['text/x-unknown', 'application/unknown', 'image/bmp', 'audio/mp3'];
+  const unsupportedMimeTypes = ['text/x-unknown', 'application/unknown', 'image/bmp'];

  // Testing general supported MIME types
  fullMimeTypesList.forEach((mimeType) => {
--- a/packages/data-provider/src/file-config.ts
+++ b/packages/data-provider/src/file-config.ts
@ -122,11 +122,27 @@ export const applicationMimeTypes =

 export const imageMimeTypes = /^image\/(jpeg|gif|png|webp|heic|heif)$/;

+export const audioMimeTypes =
+  /^audio\/(mp3|mpeg|mpeg3|wav|wave|x-wav|ogg|vorbis|mp4|x-m4a|flac|x-flac|webm)$/;
+
+export const defaultOCRMimeTypes = [
+  imageMimeTypes,
+  /^application\/pdf$/,
+  /^application\/vnd\.openxmlformats-officedocument\.(wordprocessingml\.document|presentationml\.presentation|spreadsheetml\.sheet)$/,
+  /^application\/vnd\.ms-(word|powerpoint|excel)$/,
+  /^application\/epub\+zip$/,
+];
+
+export const defaultTextMimeTypes = [textMimeTypes];
+
+export const defaultSTTMimeTypes = [audioMimeTypes];
+
 export const supportedMimeTypes = [
  textMimeTypes,
  excelMimeTypes,
  applicationMimeTypes,
  imageMimeTypes,
+  audioMimeTypes,
  /** Supported by LC Code Interpreter PAI */
  /^image\/(svg|svg\+xml)$/,
 ];
@ -169,6 +185,7 @@ export const megabyte = 1024 * 1024;
 export const mbToBytes = (mb: number): number => mb * megabyte;

 const defaultSizeLimit = mbToBytes(512);
+const defaultTokenLimit = 100000;
 const assistantsFileConfig = {
  fileLimit: 10,
  fileSizeLimit: defaultSizeLimit,
@ -192,12 +209,22 @@ export const fileConfig = {
  },
  serverFileSizeLimit: defaultSizeLimit,
  avatarSizeLimit: mbToBytes(2),
+  fileTokenLimit: defaultTokenLimit,
  clientImageResize: {
    enabled: false,
    maxWidth: 1900,
    maxHeight: 1900,
    quality: 0.92,
  },
+  ocr: {
+    supportedMimeTypes: defaultOCRMimeTypes,
+  },
+  text: {
+    supportedMimeTypes: defaultTextMimeTypes,
+  },
+  stt: {
+    supportedMimeTypes: defaultSTTMimeTypes,
+  },
  checkType: function (fileType: string, supportedTypes: RegExp[] = supportedMimeTypes) {
    return supportedTypes.some((regex) => regex.test(fileType));
  },
@ -232,6 +259,7 @@ export const fileConfigSchema = z.object({
  endpoints: z.record(endpointFileConfigSchema).optional(),
  serverFileSizeLimit: z.number().min(0).optional(),
  avatarSizeLimit: z.number().min(0).optional(),
+  fileTokenLimit: z.number().min(0).optional(),
  imageGeneration: z
    .object({
      percentage: z.number().min(0).max(100).optional(),
@ -246,6 +274,16 @@ export const fileConfigSchema = z.object({
      quality: z.number().min(0).max(1).optional(),
    })
    .optional(),
+  ocr: z
+    .object({
+      supportedMimeTypes: supportedMimeTypesSchema.optional(),
+    })
+    .optional(),
+  text: z
+    .object({
+      supportedMimeTypes: supportedMimeTypesSchema.optional(),
+    })
+    .optional(),
 });

 /** Helper function to safely convert string patterns to RegExp objects */
@ -261,7 +299,21 @@ export const convertStringsToRegex = (patterns: string[]): RegExp[] =>
  }, []);

 export function mergeFileConfig(dynamic: z.infer<typeof fileConfigSchema> | undefined): FileConfig {
-  const mergedConfig = fileConfig as FileConfig;
+  const mergedConfig: FileConfig = {
+    ...fileConfig,
+    ocr: {
+      ...fileConfig.ocr,
+      supportedMimeTypes: fileConfig.ocr?.supportedMimeTypes || [],
+    },
+    text: {
+      ...fileConfig.text,
+      supportedMimeTypes: fileConfig.text?.supportedMimeTypes || [],
+    },
+    stt: {
+      ...fileConfig.stt,
+      supportedMimeTypes: fileConfig.stt?.supportedMimeTypes || [],
+    },
+  };
  if (!dynamic) {
    return mergedConfig;
  }
@ -274,6 +326,10 @@ export function mergeFileConfig(dynamic: z.infer<typeof fileConfigSchema> | unde
    mergedConfig.avatarSizeLimit = mbToBytes(dynamic.avatarSizeLimit);
  }

+  if (dynamic.fileTokenLimit !== undefined) {
+    mergedConfig.fileTokenLimit = dynamic.fileTokenLimit;
+  }
+
  // Merge clientImageResize configuration
  if (dynamic.clientImageResize !== undefined) {
    mergedConfig.clientImageResize = {
@ -282,6 +338,26 @@ export function mergeFileConfig(dynamic: z.infer<typeof fileConfigSchema> | unde
    };
  }

+  if (dynamic.ocr !== undefined) {
+    mergedConfig.ocr = {
+      ...mergedConfig.ocr,
+      ...dynamic.ocr,
+    };
+    if (dynamic.ocr.supportedMimeTypes) {
+      mergedConfig.ocr.supportedMimeTypes = convertStringsToRegex(dynamic.ocr.supportedMimeTypes);
+    }
+  }
+
+  if (dynamic.text !== undefined) {
+    mergedConfig.text = {
+      ...mergedConfig.text,
+      ...dynamic.text,
+    };
+    if (dynamic.text.supportedMimeTypes) {
+      mergedConfig.text.supportedMimeTypes = convertStringsToRegex(dynamic.text.supportedMimeTypes);
+    }
+  }
+
  if (!dynamic.endpoints) {
    return mergedConfig;
  }
--- a/packages/data-provider/src/parameterSettings.ts
+++ b/packages/data-provider/src/parameterSettings.ts
@ -139,6 +139,18 @@ export const librechat = {
    placeholderCode: true,
    optionType: 'model',
  } as const,
+  fileTokenLimit: {
+    key: 'fileTokenLimit',
+    label: 'com_ui_file_token_limit',
+    labelCode: true,
+    description: 'com_ui_file_token_limit_desc',
+    descriptionCode: true,
+    placeholder: 'com_nav_theme_system',
+    placeholderCode: true,
+    type: 'number',
+    component: 'input',
+    columnSpan: 2,
+  } as const,
 };

 const openAIParams: Record<string, SettingDefinition> = {
@ -625,6 +637,7 @@ const googleConfig: SettingsConfiguration = [
  google.thinking,
  google.thinkingBudget,
  google.web_search,
+  librechat.fileTokenLimit,
 ];

 const googleCol1: SettingsConfiguration = [
@ -643,6 +656,7 @@ const googleCol2: SettingsConfiguration = [
  google.thinking,
  google.thinkingBudget,
  google.web_search,
+  librechat.fileTokenLimit,
 ];

 const openAI: SettingsConfiguration = [
@ -663,6 +677,7 @@ const openAI: SettingsConfiguration = [
  openAIParams.reasoning_summary,
  openAIParams.verbosity,
  openAIParams.disableStreaming,
+  librechat.fileTokenLimit,
 ];

 const openAICol1: SettingsConfiguration = [
@ -687,6 +702,7 @@ const openAICol2: SettingsConfiguration = [
  openAIParams.useResponsesApi,
  openAIParams.web_search,
  openAIParams.disableStreaming,
+  librechat.fileTokenLimit,
 ];

 const anthropicConfig: SettingsConfiguration = [
@ -702,6 +718,7 @@ const anthropicConfig: SettingsConfiguration = [
  anthropic.thinking,
  anthropic.thinkingBudget,
  anthropic.web_search,
+  librechat.fileTokenLimit,
 ];

 const anthropicCol1: SettingsConfiguration = [
@ -721,6 +738,7 @@ const anthropicCol2: SettingsConfiguration = [
  anthropic.thinking,
  anthropic.thinkingBudget,
  anthropic.web_search,
+  librechat.fileTokenLimit,
 ];

 const bedrockAnthropic: SettingsConfiguration = [
@ -736,6 +754,7 @@ const bedrockAnthropic: SettingsConfiguration = [
  bedrock.region,
  anthropic.thinking,
  anthropic.thinkingBudget,
+  librechat.fileTokenLimit,
 ];

 const bedrockMistral: SettingsConfiguration = [
@ -747,6 +766,7 @@ const bedrockMistral: SettingsConfiguration = [
  mistral.topP,
  librechat.resendFiles,
  bedrock.region,
+  librechat.fileTokenLimit,
 ];

 const bedrockCohere: SettingsConfiguration = [
@ -758,6 +778,7 @@ const bedrockCohere: SettingsConfiguration = [
  cohere.topP,
  librechat.resendFiles,
  bedrock.region,
+  librechat.fileTokenLimit,
 ];

 const bedrockGeneral: SettingsConfiguration = [
@ -768,6 +789,7 @@ const bedrockGeneral: SettingsConfiguration = [
  meta.topP,
  librechat.resendFiles,
  bedrock.region,
+  librechat.fileTokenLimit,
 ];

 const bedrockAnthropicCol1: SettingsConfiguration = [
@ -787,6 +809,7 @@ const bedrockAnthropicCol2: SettingsConfiguration = [
  bedrock.region,
  anthropic.thinking,
  anthropic.thinkingBudget,
+  librechat.fileTokenLimit,
 ];

 const bedrockMistralCol1: SettingsConfiguration = [
@ -802,6 +825,7 @@ const bedrockMistralCol2: SettingsConfiguration = [
  mistral.topP,
  librechat.resendFiles,
  bedrock.region,
+  librechat.fileTokenLimit,
 ];

 const bedrockCohereCol1: SettingsConfiguration = [
@ -817,6 +841,7 @@ const bedrockCohereCol2: SettingsConfiguration = [
  cohere.topP,
  librechat.resendFiles,
  bedrock.region,
+  librechat.fileTokenLimit,
 ];

 const bedrockGeneralCol1: SettingsConfiguration = [
@ -831,6 +856,7 @@ const bedrockGeneralCol2: SettingsConfiguration = [
  meta.topP,
  librechat.resendFiles,
  bedrock.region,
+  librechat.fileTokenLimit,
 ];

 export const paramSettings: Record<string, SettingsConfiguration | undefined> = {
--- a/packages/data-provider/src/schemas.ts
+++ b/packages/data-provider/src/schemas.ts
@ -680,6 +680,8 @@ export const tConversationSchema = z.object({
  iconURL: z.string().nullable().optional(),
  /* temporary chat */
  expiredAt: z.string().nullable().optional(),
+  /* file token limits */
+  fileTokenLimit: coerceNumber.optional(),
  /** @deprecated */
  resendImages: z.boolean().optional(),
  /** @deprecated */
@ -794,6 +796,8 @@ export const tQueryParamsSchema = tConversationSchema
     * https://platform.openai.com/docs/api-reference/runs/createRun#runs-createrun-instructions
     * */
    instructions: true,
+    /** @endpoints openAI, google, anthropic */
+    fileTokenLimit: true,
  })
  .merge(
    z.object({
@ -850,6 +854,7 @@ export const googleBaseSchema = tConversationSchema.pick({
  thinking: true,
  thinkingBudget: true,
  web_search: true,
+  fileTokenLimit: true,
  iconURL: true,
  greeting: true,
  spec: true,
@ -1101,6 +1106,7 @@ export const openAIBaseSchema = tConversationSchema.pick({
  useResponsesApi: true,
  web_search: true,
  disableStreaming: true,
+  fileTokenLimit: true,
 });

 export const openAISchema = openAIBaseSchema
@ -1145,6 +1151,7 @@ export const anthropicBaseSchema = tConversationSchema.pick({
  spec: true,
  maxContextTokens: true,
  web_search: true,
+  fileTokenLimit: true,
 });

 export const anthropicSchema = anthropicBaseSchema
--- a/packages/data-provider/src/types/files.ts
+++ b/packages/data-provider/src/types/files.ts
@ -47,6 +47,7 @@ export type FileConfig = {
  endpoints: {
    [key: string]: EndpointFileConfig;
  };
+  fileTokenLimit?: number;
  serverFileSizeLimit?: number;
  avatarSizeLimit?: number;
  clientImageResize?: {
@ -55,6 +56,39 @@ export type FileConfig = {
    maxHeight?: number;
    quality?: number;
  };
+  ocr?: {
+    supportedMimeTypes?: RegExp[];
+  };
+  text?: {
+    supportedMimeTypes?: RegExp[];
+  };
+  stt?: {
+    supportedMimeTypes?: RegExp[];
+  };
+  checkType?: (fileType: string, supportedTypes: RegExp[]) => boolean;
+};
+
+export type FileConfigInput = {
+  endpoints?: {
+    [key: string]: EndpointFileConfig;
+  };
+  serverFileSizeLimit?: number;
+  avatarSizeLimit?: number;
+  clientImageResize?: {
+    enabled?: boolean;
+    maxWidth?: number;
+    maxHeight?: number;
+    quality?: number;
+  };
+  ocr?: {
+    supportedMimeTypes?: string[];
+  };
+  text?: {
+    supportedMimeTypes?: string[];
+  };
+  stt?: {
+    supportedMimeTypes?: string[];
+  };
  checkType?: (fileType: string, supportedTypes: RegExp[]) => boolean;
 };

--- a/packages/data-schemas/src/schema/defaults.ts
+++ b/packages/data-schemas/src/schema/defaults.ts
@ -141,6 +141,9 @@ export const conversationPreset = {
  disableStreaming: {
    type: Boolean,
  },
+  fileTokenLimit: {
+    type: Number,
+  },
  /** Reasoning models only */
  reasoning_effort: {
    type: String,
--- a/packages/data-schemas/src/schema/preset.ts
+++ b/packages/data-schemas/src/schema/preset.ts
@ -51,6 +51,7 @@ export interface IPreset extends Document {
  useResponsesApi?: boolean;
  web_search?: boolean;
  disableStreaming?: boolean;
+  fileTokenLimit?: number;
  // end of additional fields
  agentOptions?: unknown;
 }
--- a/packages/data-schemas/src/types/convo.ts
+++ b/packages/data-schemas/src/types/convo.ts
@ -50,6 +50,7 @@ export interface IConversation extends Document {
  useResponsesApi?: boolean;
  web_search?: boolean;
  disableStreaming?: boolean;
+  fileTokenLimit?: number;
  // Additional fields
  files?: string[];
  expiredAt?: Date;