danny-avila/LibreChat · warning · Error

Extracted text from "${file.originalname}" exceeds the 15MB

Error message

Extracted text from "${file.originalname}" exceeds the 15MB storage limit (${Math.round(textBytes / megabyte)}MB). Try a shorter document.

What it means

Thrown by the inner createTextFile helper when the UTF-8 byte length of extracted text (from OCR, RAG, STT, or document parsing) exceeds 15 MB. This is a hard storage cap on extracted text persisted to the file/text field; it prevents one giant document from blowing up the DB row or downstream token budgets. The original filename and the computed MB are included.

Source

Thrown at api/server/services/Files/process.js:758

    if (!isFileSearchEnabled) {
      throw new Error('File search is not enabled for Agents');
    }
    // Note: File search processing continues to dual storage logic below
  } else if (tool_resource === EToolResources.context) {
    const { file_id, temp_file_id = null } = metadata;

    /**
     * @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 textBytes = Buffer.byteLength(text, 'utf8');
      if (textBytes > 15 * megabyte) {
        throw new Error(
          `Extracted text from "${file.originalname}" exceeds the 15MB storage limit (${Math.round(textBytes / megabyte)}MB). Try a shorter document.`,
        );
      }
      const retentionExpiry = await getAgentFileRetentionExpiry({
        req,
        messageAttachment,
        tool_resource,
      });
      const fileInfo = {
        ...removeNullishValues({
          text,
          bytes,
          file_id,
          temp_file_id,
          user: req.user.id,
          type,
          filepath: filepath ?? file.path,
          source: FileSources.text,

View on GitHub (pinned to 5ff282f900)

Solutions

Upload a shorter document — split the source PDF into chunks under the limit.
For long audio, trim or segment before STT so the transcript stays under 15 MB.
If extraction is producing inflated/garbage text, investigate the OCR/RAG parser — duplicated output suggests a parser bug.
Do not raise the 15 MB constant casually; it also bounds DB row size and downstream LLM token cost.

Defensive patterns

Strategy: validation

Validate before calling

const TEXT_LIMIT = 15 * 1024 * 1024;
function assertExtractedTextSize(text, filename) {
  const bytes = Buffer.byteLength(text, 'utf8');
  if (bytes > TEXT_LIMIT) {
    throw new Error(`Extracted text from ${filename} is ${Math.round(bytes / (1024*1024))}MB; limit is 15MB`);
  }
}

Try / catch

try { await processAgentFileUpload(params); }
catch (e) {
  if (/exceeds the 15MB storage limit/.test(e.message)) return res.status(413).json({ error: 'Document too large to index; please split it.' });
  throw e;
}

Prevention

Split large documents (books, long reports) into sub-15 MB chunks before upload.
For long audio, segment before STT to bound transcript size.
Investigate inflated/garbage text from OCR/RAG — duplication suggests a parser bug.

When it happens

Trigger: A document whose extracted text crosses 15 MB — e.g., a 2,000-page PDF, a large transcribed audio file, or an OCR'd scan of a book. The text length is computed via `Buffer.byteLength(text, 'utf8')`, not the source file size.

Common situations: Users uploading entire books or massive reports; OCR of a high-page-count scan; a long webinar's STT output; a malformed text extractor that emits duplicated/garbage text inflating byte count.

Related errors

AI-assisted analysis of danny-avila/LibreChat@5ff282f900 (2026-08-12). Data as JSON: /api/errors/bef3f0fb75548c4a. Report an issue: GitHub.