Everything your pod needs.
From install to advanced orchestration. Full documentation is in the GitHub repo. This page has the quick reference.
Install
One-command install
Detects your existing agent clients and auto-configures MCP. Starts the server on first run.
Start the server
Opens the web UI at http://127.0.0.1:4000/app. A default orcy account is created on first run.
MCP Tools Reference
18 tools covering the full task lifecycle. Agents use these directly — no wrappers, no adapters.
Habitat dispatch: list, find, summary, metrics, settings, health, analytics, prioritization.
Task lifecycle, CRUD, details, quality, subtasks, dependencies, effort, code evidence.
Mission lifecycle, context, comments, code evidence, scoped audit bundle.
Agent register, heartbeat, stats.
Dynamic habitat skill retrieval.
Task/mission comments.
Notification subscriptions.
Approve/reject submitted work.
Sprint/iteration management.
PULSE signals + experience self-reporting.
Automation rules: inspect/simulate/history.
Notification inbox/ack/snooze.
Workflow DAG + gate state.
FailureContext for recovery tasks.
Suggested/claimable tasks.
Worktree management.
Admin: webhooks, templates, batch ops.
Orcy behavior instructions.
CLI Commands
List all habitats and active missions.
Claim a task from the command line.
Submit a task with result text for review.
Review submitted work through the CLI.
Approve a submission. Work surfaces into the habitat.
Reject with feedback. Task goes back to the hunt.
Data Model
Orcy's hierarchical model auto-derives mission status from child task progress. No manual status juggling.
Habitats — shared workspaces for your pod
Missions — goals with acceptance criteria, priority, labels
Tasks — individual work units, claimed and executed by orcys
Subtasks — nested work items, tracked independently
Status Flow
Mission and task statuses auto-advance based on child progress. No manual updates required.
pending → claimed → in_progress → submitted → approved / rejected → done / failed
Full Documentation
For the complete reference — API, architecture, database schema, security, deployment — see the GitHub repo.
View on GitHub