Documentation

Connect an agent, add the keys it needs, and see every call it makes.

Make your first call

Save a service key, create an agent, connect it over MCP, make its first call and answer its first approval.

Add service keys, connect an MCP client, or send plain HTTP through the gateway.

Keys and access
Add service keys and decide which agents may use them.
MCP
Connect an MCP client with a token or by signing in, and route remote MCP servers through Sallyport.
HTTP proxy
Call a service over plain HTTP with an agent token, from scripts and CI jobs.

Approvals
Answer a waiting call, choose how long a permission lasts, and handle access requests.
Troubleshooting
Find where a call stopped from its answer, its request ID and the Activity screen.

The console API, the service catalog, the journal, SSH commands, security and limits.

REST API
Manage agents, keys and access from your own scripts with a console key.
Service recipes
How a catalog service is described: its hosts, its key fields and how its calls are classed.
Journal export
Download the signed journal as JSON Lines, or send journal events to your own webhook.
Email
Let an agent send from your mailbox, search and read it, and sort its mail, while Sallyport keeps the password.
SSH commands
Let an agent run commands on a public server while Sallyport holds the private key.
Files over SFTP
Let an agent list, read, write and move files on your server while Sallyport holds the private key.
S3 storage
Connect a bucket of any S3-compatible storage and let agents list, read, write and delete its files.
Security model
Where keys live, how accounts stay apart, what the network guard refuses and how the journal is signed.
Limits
Every limit on calls, sizes and time, where to change it and what happens past it.