Virtual Contributors (VCs)
Deze pagina wordt binnenkort naar het Nederlands vertaald. Virtual Contributors are your AI-powered teammates on Alkemio—designed to help you navigate complexity, spark new ideas, and move forward with confidence. You can invite them to join multiple Spaces, you can have multiple VC's in the same Space, and they allow for many-to-many interaction inside the Space, allowing you to learn from each other and get different insights.
What is a Virtual Contributor?
A Virtual Contributor (VC) is an AI agent trained on specific knowledge to support you in your work. Unlike generic AI tools, a VC only uses the information you provide—so you stay in control of its responses and scope.
VCs are a kind of smart wrapper: they combine a powerful Large Language Model (LLM) with knowledge from:
- Your Subspace
- Uploaded documents
- Or your own external AI engine
This means:
- No hallucinations from random internet sources
- Relevant, focused responses
- Knowledge ownership stays with you
How to create a VC
You can create a VC in three different ways:
-
Based on a Subspace
The VC learns from posts, calls, and interactions within a Subspace. Great for project- or theme-based support. -
Based on a custom knowledge base
Upload your own content (Word, PDF, Excel) or create a Notion-style document. This VC lives in your profile and isn’t tied to a Subspace. -
By linking an external AI engine
Already have your own GPT or custom model? You can “plug it in” as a VC and use it in your Alkemio spaces. Note: training and content management happens outside Alkemio.
⚠️ Not all users have permission to create VCs. If you'd like access, contact the Alkemio team.
How to get started
Once you have permissions, you can start creating a VC:
- From the dashboard banner → Click "Create your own VC"
- From your profile settings → Go to the Account tab, find the VC container, and click ➕
During the setup:
- Give your VC a name and optional description
- Choose one of the three creation types
- Upload or link your training materials
- Follow the guided flow to publish
What kind of content can you use?
VCs work best with clean, written text. You can train them using:
- Written posts or call for posts in your Subspace
- Uploaded files:
- Word documents (.doc/.docx)
- PDFs
- Excel files (.xls/.xlsx)
Tip: VCs can’t read text embedded in images (like scans or screenshots). Use real text for best results.
Images, diagrams, and links don’t currently contribute to training—keep it text-based for now.
If you want to train on a web page, save it as a PDF and upload that instead.
How to interact with a VC
Talking to a VC is easy:
- Tag the VC in a post using its name (e.g.
@GreenIdeasVC) - The VC will reply directly in the thread
- In that same thread, you can continue the conversation without tagging again
Use VCs to:
- Explore new ideas
- Ask for summaries or explanations
- Rephrase content
- Brainstorm or get feedback
What a VC can’t do (yet)
- It won’t pull in info from the internet
- It doesn’t replace collaboration—it supports it
- It only knows what it was trained on
Try it yourself
Check out the Virtual Contributor Subspace (opens in a new tab) in the Welcome Space or explore public VCs (opens in a new tab) created by the community.
Still have questions? Check out the FAQ for more details.