What Rykon AI does, and where to find each part
AI agents that do bounded work inside rules you set, with a person deciding anything above their level and a record of every action.
Updated 1 min read
Rykon AI gives your company AI agents that do real work. You describe a task; simple answers come straight back, and anything that runs code, a shell or files gets a fresh sandbox of its own; every model call goes through our AI gateway on your account’s own metered key; anything the agent wants to do above its level waits for a person, with the exact action shown; and everything that happened is kept on a record nobody can edit.
Everything is managed in the console at app.rykon.ai. Your own code reaches the same things through the API.
The console menu
| Menu item | What it is for |
|---|---|
| Home | What needs you (approvals waiting, runs that failed), what your agents made, and what ran. |
| Agents | Your agents, and a gallery of templates by department to start one from. |
| Runs | Start a task, see every run, open one to watch it live or read its transcript and results. |
| Approvals | Actions waiting for a person, with the exact action, ready to approve, edit or reject. |
| Usage | Model cost, tokens and calls, per day and per agent. |
| Team | The people on the account and their roles. |
| API keys | Keys for your own code, each with only the scopes you give it. |
| Audit log | Who did what to the account, and the agents’ record of what they did. |
| Settings | The account name, the account kill switch, your sign-in and the appearance. |
Who can change what
Every person on an account is an owner, an admin, a member or a viewer. Members work with agents and runs; viewers read; owners and admins also manage people and keys and may let an agent act with more autonomy. See Roles and what each can do.