Create an Agent

Creating a new agent

Go to Console → Agents and choose New agent. With no agents yet, the Console home asks the same question.
Describe the work and choose Draft with AI, pick a ready agent from the four shelves, or choose Start blank at the end of the page.
Check the details: choose Change something on a draft, or open a ready agent’s card. Use this agent skips this step.
Choose Create and try it. The agent is created and its first chat opens in the same tab.

While Draft with AI works, the page shows each step it takes. Cancel stops it, and so does leaving the page; nothing is created until you choose Create and try it.

Opening an agent shows everything about it on one page. A rail down the left moves you between its sections (instructions, model, Knowledge, Skills, Integrations, autonomy, Capabilities, voice, Channels, access, its computer, and its record), and one save bar at the bottom tells you what you have changed and saves it in one go. Sections that apply the moment you touch them (Skills, Channels, access) say so. A Customer-facing agent has no autonomy or computer section.

Try it opens beside the settings, so you can send the agent a message and read its answer without leaving the page. It talks to the saved agent, and says so while you have unsaved changes.

The first time you open a new agent, its checklist (Getting this agent ready) opens by itself: instructions, Integrations, Connections, Knowledge (optional) and a first try. The ring at the top shows how many are done, and goes away once they all are.

History lists recent versions of the configuration — who changed what, and when — and can put an earlier version back into the editor. Nothing is written until you save, so a restore is something you look at first. Edits made in one sitting are kept as a single entry rather than one per pause.

Basic info

Field	Description
Name	What your agent is called. Change it later with Rename agent on the agent’s page.
What should this agent do?	One line that becomes the description your team sees.
Who will talk to it?	Your team (an Internal agent) or Your customers (a Customer-facing agent). Chosen at creation and fixed after. An Internal agent can also switch on Let people outside your team join its chats. See Agent Types →.

Instructions

The instructions define the agent’s behavior. Write them in plain English.

See Writing Agent Instructions → for detailed guidance and examples.

Capabilities

Capabilities are what the agent can do on its own, built into Auteryn. Switch on only the ones this agent needs:

Capability	What it enables
Web search	Search the live web
Location & maps	Ground answers in Google Maps places data
Browser	Navigate and act on web pages, including vision-based clicks and typing (Computer Use). On by default
Coding	Repo-aware code editing with developer tooling on the agent’s computer. Paid plans
Presentations, Diagrams, Image generation, Audio generation, Video generation	Create that kind of output. Video generation needs a paid plan
Memory	Remember across conversations. On by default
File upload, Google Drive files, YouTube links, Voice input, Camera, Screen share	What people can attach or capture in the chat composer

An Internal agent always has its own computer, a sandbox with a terminal and files, so there is no switch for it. Customer-facing agents don’t get a sandbox or browser. Voice has its own section. See Sandboxes → and Computer Use & Browser →.

Integrations

Select which Integrations the agent can use, and the connection it acts through. An owner or admin can also add an Integration that isn’t connected yet without leaving the agent’s page.

On a team org, each attached connection shows who it acts as (“Acts as: you”, “Acts as: <name>” or “Acts as: team (shared)”). You can only attach a teammate’s personal connection if they’ve shared it with you.

Acts as each user — for sign-in based integrations like Notion or Jira on an internal agent, you can flip the binding to act as each teammate’s own account: everyone connects their own login with the Connect my account card on the integration’s panel (it shows when an agent that uses the integration acts as each person), and the agent uses the runner’s account. Choose what happens for teammates who haven’t connected theirs: quietly use the team connection (default) or ask them to connect their own first. In chat, a run that needs your own account shows Connect your <service> account, which opens the Console to connect it. Computer-side integrations (GitHub, Google Workspace, cloud CLIs) always act as one team identity, and customer conversations always use the team connection.

Knowledge

Attach one Knowledge collection. When attached, the agent will automatically search it during conversations.

See Create Knowledge →.

Skills

Attach skills from the system library or your org’s custom skills. Skills are workflow templates the agent can follow for specific task types.

See Using Skills → and Skill Library →.

Model

Choose which model thinks for the agent, and How hard it thinks: Auto, Fast, or Deep. Internal agents default to Auto, where Auteryn picks per message; Customer-facing agents always run Fast. The Free plan has no model choice: its agent runs on Auto. See Execution Modes →.

Autonomy

How much an Internal agent may do before it checks with a person, and rules for specific actions. See Autonomy & Approvals →.

Deleting an agent

Deleting is an owner/admin action — members can’t delete agents (see Roles & permissions).

Delete any Loops that use the agent; an agent with Loops can’t be deleted
Open the agent and go to Danger zone in the left rail
Under Delete permanently, choose Delete
Confirm with Delete agent