Vello offers the capability to train your AI Personas using a vast array of documents, acting as a knowledge base which they can reference and learn from. This enhances their ability to provide contextually accurate and informed responses.

​
Training Process

To train a persona with files, follow these steps:
  1. Accessing File Upload: In the chat with your chosen persona, click on the document icon or drag and drop your files into the chat interface. The personas will store these files in memory, referring to them as needed.
  2. Supported File Formats: You can upload various types of documents. Currently, Vello supports the following formats:
    • C (text/x-c)
    • C++ (text/x-c++)
    • CSV (application/csv)
    • DOCX (application/vnd.openxmlformats-officedocument.wordprocessingml.document)
    • HTML (text/html)
    • Java (text/x-java)
    • JSON (application/json)
    • Markdown (text/markdown)
    • PDF (application/pdf)
    • PHP (text/x-php)
    • PPTX (application/vnd.openxmlformats-officedocument.presentationml.presentation)
    • Python (text/x-python, text/x-script.python)
    • Ruby (text/x-ruby)
    • LaTeX (text/x-tex)
    • Plain Text (text/plain)
  3. Knowledge Integration: Once your files are uploaded, the persona can reference this content to respond to queries. This is particularly useful for complex tasks where domain-specific knowledge is essential.
  4. Ongoing Training: You can continue to add to your persona’s knowledge base by uploading additional documents as they become relevant to your team’s work.
  5. Future Upgrades: Vello is continuously working on enhancing its features. An upcoming release will include website ingest support, allowing personas to incorporate information directly from web content.
By effectively using this feature, your personas become more attuned to the specifics of your projects and industry, offering tailored assistance that leverages the collective wisdom of your organization.