head to https://pingu.audn.ai and login or register
head to https://pingu.audn.ai/mcp
Click deploy MCP button
It requires 1 day free trial to Audn Premium
PS: You might need to do checkout twice if your premium is not activated on the first time or you used multiple accounts.
I recommend clearing cache and history and logging in again on a fresh browser session and starting the trial there.
3. head to https://pingu.audn.ai/mcp and press deploy MCP button again
you will see this UI if you activated your trial successfully.
4. Enter your classic legacy github personal access token ( if you don't have one create by following these steps:
4.a. head to https://github.com/settings/tokens and generate a personal access token. The token should have read and write access to your repositories and only used to fork a public or private repo you will share and will build on the forked repo which will be placed in your account privately.
)
The agent writes code so make sure it can write code in your repos.
5. Find a github MCP url or search for one existing in our internal MCP marketplace. Leave the subdirectory empty if MCP implementation is on the root if not please give subdirectory.
You can also search for MCPs on github
https://github.com/search?q=cybersecurity+MCP&type=repositories
One good example is
https://github.com/cyproxio/mcp-for-security
and any subdirectories it has such as
https://github.com/cyproxio/mcp-for-security/tree/main/amass-mcp
in that case enter amass-mcp as subdirectory path
Make sure to select the platform you want to deploy:
Use Vercel for serverless friendly mcps
Use Render for long living docker supporting mcps that requires specific ports and configurations
Both require your personal access token. They are not stored and discarded after the deployment.
Important! Use hopx.ai MCP if you want to deploy an MCP server with sudo access requirement. Because only hopx.ai vm sandbox give you that option. ( Some networking MCPs require this)
Press start AI conversion:
You can save progress if it took so long and you can continue later. It might take around 30 to 60 minutes.
When you save progress it appears on the same page like this, you can press edit and then press continue or tell the agent a different direction.
If it asks for feedback try to get error logs from the platform you tried to deploy and give agent a feedback.
After deployment is completed it should be added automatically to your account if it's not added you can use this UI to add your publicly exposed MCP.
You can add any MCP you have and they will continue working even after trial because adding your own MCPs to pingu is enabled for free-tier users as well.
After your MCP is successfully added it appears like this
You can press test to test the mcp connection. If it lists number of tools in green it's working. Just start a new chat and your MCP will be autoconnected on a new chat.
