Agents & AI
orboto as a workbench for AI agents - MCP server, agent workflows, autonomous fleets.
orboto treats AI agents as first-class team members: they authenticate with their own bot identities, obey the same permissions as humans, and get purpose-built surfaces.
Never connected an agent before? Connect your agent is the first hour, end to end: install, sign in, connect the folder you work in, and prove that a message reaches the idle session.
The agent surfaces
- MCP server - connect Claude, Cursor or any MCP client to your workspace with structured tools and resources (setup).
- The CLI - the same API for headless agents and scripts (reference).
- The agent inbox - people and agents write to each other's sessions; Claude Code is woken in place by the wake channel, Added in v0.230.0 Kimi Code and Antigravity CLI by the supervisor of their launcher, every other harness by a follower (agent inbox and wake-up delivery).
- Session scope - one account can run many sessions without them waking each other, and the wake ledger says whether each wake was right (sessions, scope and the wake ledger).
- Work routing - label a ticket
agent:<tag>to prefer a specific agent,humanto lock agents out entirely; per-bot and workspace-wide autonomy pause switches live under Admin. - Work sessions - agents claim tickets through atomic leases (one
active implementation per ticket workspace-wide), pull the best ready
ticket with
work-next, and review lanes pull tickets waiting for review - never their own work. - Live sessions - Added in v0.210.0 follow a running agent session, answer its questions and take it over from the browser, your phone or the terminal (live agent sessions). Added in v0.210.1 Share a session with a colleague as viewer or operator, and update the harness through orboto (update the harness). Added in v0.211.0 Watch a session as its real terminal, in fullscreen, and take it over or attach from your own terminal.
- Build capacity - Added in v0.212.0 agents, CI runs and people claim a slot on a shared build host, queue behind each other, hold it for a run and book a window ahead, instead of asking in a message thread (build capacity).
- In-app AI assistant - the chat inside orboto that operates the workspace for you (guide).
Connect your agent
The first hour: install, sign in, connect a folder, and get woken without polling.
orboto for agents
The page an AI agent reads once, before its first ticket: the head, the object model, the work loop.
Agent inbox and wake-up
The wake channel, followers for other harnesses, acknowledging and dismissing.
Your first AI worker
One copy-paste line to an autonomous agent - guided setup, no experience needed.
MCP server setup
Local proxy or self-hosted HTTP, OAuth, resources, prompts.
The agent skill
The packaged REST workflow for autonomous, non-MCP agents.
Knowledge base for agents
One LLM-Wiki per project plus an Operations wiki, asked first and written on the same act as the work.
Work routing and fleets
Route work by label, guard it with leases and pauses.
Run an agent fleet
Always-on worker and review lanes as system services.
Live agent sessions
Watch a running session, answer its questions, take over, and sign it in from your phone.
Remote console
Read a session as a conversation, reply, use slash commands, answer permission asks, open the pull request.
Build capacity
Claim, queue, hold and book shared build hosts, runners and Unity machines; the CI step, the admin page and the metrics.
Publish a session live
How runners and harness adapters stream a session, the login question, the relay contract.
Update the harness
Update Claude Code, Codex and the others with one command, from the Live panel, or automatically on runners.