File Uploads

File upload is the fastest way to get content into CoAgentor. Upload any supported document and it becomes available for your agents to reference during live meetings.

→ Upload files →

Supported file types#

| Format | Extension | |---|---| | PDF | .pdf | | Word Document | .docx | | Excel Spreadsheet | .xlsx | | CSV | .csv | | Plain Text | .txt | | Markdown | .md | | HTML | .html |

More formats may be supported — if you have a specific format requirement not listed, contact support.

Uploading a file#

1. Go to Data Sources →
2. Click Upload File
3. Select one or more files from your computer
4. Files will appear in the list immediately with Pending status
5. Processing begins automatically — status will update to Completed within a few minutes

You can upload multiple files at once. Each file is processed independently.

What happens after upload#

CoAgentor processes each file through the following pipeline:

1. Text extraction — format-specific text extraction (PDF pages, Word paragraphs, spreadsheet cells, etc.)
2. Chunking — the extracted text is split into 512-token segments with 50-token overlap. Overlap ensures that content spanning a chunk boundary isn't lost.
3. Embedding — each chunk is embedded as a semantic vector using an AI embedding model
4. Indexing — chunks are stored and ready for retrieval during meetings

You don't need to do anything during processing. Once status shows Completed, the file is ready to add to a context.

Adding a file to a context#

Uploading a file doesn't automatically make it available to your agents. You need to add it to a Context and assign that context a scope.

→ Learn about Contexts and Scopes →

Contexts let you control which files are used for which agents or meetings — so a sales agent only references your sales playbook, not your HR documentation.

Deleting a file#

To delete a file:

1. Go to Data Sources →
2. Click the options menu on the file row
3. Select Delete

Deleting a file removes it from all contexts it belongs to and permanently removes the stored chunks. This cannot be undone.

File size and limits#

• Supported encodings: UTF-8 (for text files). Binary formats (PDF, DOCX, XLSX) are handled automatically.
• Plan-level limits also apply to how many chunks can be retrieved per meeting evaluation.

Best practices#

✅ Name files descriptively — the filename helps you identify the file later in contexts (e.g. sales-playbook-Q1-2025.pdf rather than doc1.pdf)

✅ Use structured documents — well-formatted documents with headings and clear sections produce better chunk boundaries and more accurate retrieval

✅ Keep documents focused — a tight, purpose-specific document (e.g. "Objection handling for enterprise accounts") retrieves better than a large, mixed document

✅ Update files when content changes — delete the old file and re-upload the updated version. CoAgentor doesn't auto-sync uploaded files.

❌ Don't upload scanned PDFs with poor OCR — text extraction quality depends on readable text content, not images of text

Troubleshooting#

Processing stuck on Pending — Processing typically starts within 60 seconds. If a file stays in Pending for more than 5 minutes, try re-uploading.

Processing Failed — Click the error badge to see the failure message. Common causes: unsupported file encoding, password-protected PDFs, or corrupt files.

My agent isn't using the file content — Make sure the file is in a context and that context has a scope covering the current agent or meeting. Check Contexts → to verify.

→ Set up contexts →