If you’ve seen how tools connect to AI products like ChatGPT or Claude, this is the same kind of upgrade now available in Vera.
Vera now supports MCP (Model Context Protocol), which gives AI assistants a structured way to connect to product actions. In practice, that means the assistant can do more than generate answers. It can now interact with Vera through defined operations that help users move work forward in a more reliable and controlled way.
With MCP support, Vera becomes more than a conversational layer. It becomes an action layer.
What This Means in Vera
MCP gives the assistant a direct path into real Vera workflows. That means the assistant can now support a much wider range of structured actions across questionnaire operations, helping users go beyond asking for help and into actually getting tasks done.
With Vera’s MCP support, you can now work across actions such as:
- discovering companies, labels, and users
- searching and managing past answers
- answering questions with supporting sources
- listing, adding, updating, deleting, and regenerating questions
- creating, updating, deleting, and exporting questionnaires
- creating, updating, reordering, and deleting sections
- retrying extraction when a questionnaire needs another pass
This makes the assistant far more useful in day-to-day work, especially for teams handling security questionnaires, compliance workflows, and other structured review processes.
Why MCP Matters
This is a meaningful shift in how Vera can be used. Instead of relying only on open-ended prompts, MCP gives the assistant access to clearly defined operations with clearer intent and more predictable outcomes. It makes Vera feel less like a chat box beside the workflow and more like an assistant that can actively operate within it.
.png)
Built with Guardrails
Some actions should remain confirmation-based.
For more sensitive flows, such as deleting content, creating questionnaires, or triggering exports, Vera may return a guided UI path instead of executing the action immediately. This gives users a clear review point before completing higher-impact operations.
That balance is intentional. MCP makes Vera more capable, while still keeping the right safeguards in place.
How to Set Up Vera MCP
1. Add the Vera MCP server to your MCP-compatible client
To connect to Vera’s MCP server,
- For ChatGPT follow this guide https://developers.openai.com/apps-sdk/deploy/connect-chatgpt
- For Claude.ai Web/Desktop app (Reference):
- Navigate to Customize > Connectors.
- Click the “+” button next to Connectors.
- Select “Add custom connector.”
- Enter the connector's name and URL (
https://beta-api.privasee.io/mcp). - Follow the authentication with Vera and start using.
- For Claude Code (Reference):
- Use the following command:
claude mcp add --transport http vera https://beta-api.privasee.io/mcp
- In Claude Code, type
/mcp, select "vera" from the list, and click Authenticate to log in with your Vera account.
- In Claude Code, type
- For any other MCP Clients use the connection endpoint
https://beta-api.privasee.io/mcpand choose OAuth Authentication.
2. Authenticate using your normal Vera access flow
Your MCP client should connect using the authentication flow your Vera environment expects. Once connected, the assistant can access the structured Vera actions available to your workspace.
3. Verify Installation and tool access
A good way to confirm setup is working is to start with lightweight actions such as:
- list companies
- list labels
- list users
- list questionnaires
If those work, the connection is live and the assistant can access your workspace context.
.png)
4. Move on to real workflow actions
Once connected, you can begin using MCP for more practical actions such as:
- updating questions
- reordering sections
- searching and managing past answers
- retrying extraction
- exporting questionnaires
A New Layer of Vera Is Now Live
This is more than just a new integration. It is a foundational step toward more agentic workflows inside Vera.
By introducing MCP, Vera gains a structured action layer that allows the assistant to do more than respond. It can now take part in real workflow execution across questionnaire and knowledge operations in a way that is more reliable, more controlled, and more useful in day-to-day work. The assistant is no longer just there to help with the work. It is becoming a more active part of how the work gets done.