Platform Features
Everything You Need to Orchestrate AI Agents
From live terminal streaming to inter-agent contracts, permission governance to mobile notifications. Purpose-built for teams running AI coding agents at scale.
Task Management & Workflow Engine
Purpose-built task system with agent-native state machines, structured specs, and dependency orchestration.
Visual Kanban Board
8-column board tracking tasks from pending through planning, execution, review, and completion.
Smart Auto-Assignment
Capability matching scores your servers against task requirements and assigns the best fit automatically.
Plan Mode
Require agents to submit execution plans for human review before they start coding. Approve or request revisions.
Structured Task Specs
Define requirements with MoSCoW prioritization, constraints, and expected outputs.
Task Dependencies
Chain tasks together with typed dependencies (blocks, input, related) so downstream work auto-starts when upstream completes.
Task Notes
Send notes to running agents. Urgent notes interrupt immediately, normal notes wait for an idle moment.
Bulk Operations
Archive, unarchive, or delete multiple tasks at once from the list view.
Live Terminal Streaming
Watch your AI agents work in real time — exactly like sitting at the terminal, but from anywhere.
Real-Time Terminal
Full xterm.js terminal in your browser with live PTY streaming via WebSocket.
Multi-Tab Workspace
Per-project tabbed terminal view. Switch between agents with keyboard shortcuts (Alt+1-9, Ctrl+Tab).
Interactive Control
Type into running sessions, send keyboard shortcuts, resize terminals on the fly.
Text View Mode
Toggle to plain-text reading mode for easier scanning of agent output.
Full Scrollback
Retrieve complete terminal history on demand — never lose context.
Popout Windows
Open any session in a standalone window, synced back to the main workbench.
Session Recovery
If an agent daemon restarts, orphaned sessions are automatically rediscovered and reattached.
Inter-Agent Orchestration
Coordinate multiple AI agents working on related tasks across different repositories and servers.
Contract-Based Handoffs
Define what output one agent must produce for the next agent to consume. Contracts are validated automatically.
Task Cascading
When an upstream task completes, downstream tasks are automatically unblocked and assigned.
Dependency Graph
Visualize and manage task dependencies with blocks, input, and related dependency types.
Cross-Repo Coordination
Agents on different servers working on different repos, all orchestrated from one dashboard.
Human-in-the-Loop
AI agents that know when to ask for help. You respond from the dashboard. Work continues.
Agent Help Requests
Agents can request human intervention (help, questions, or approval) when they are stuck.
Intervention Queue
Pending requests surface in the dashboard with context and the agent's question.
One-Click Response
Type your answer, approve, or reject. The response is injected directly into the agent's terminal.
Plan Approval Gates
Require agents to submit plans for review before execution. Approve, reject, or request revisions.
Server Management & Onboarding
Connect your servers in minutes. Three tiers to match your setup.
Full Agent Daemon
Persistent daemon with real-time WebSocket, session management, file transfer, and SSH tunneling. Linux, macOS, Windows.
Skill-Only Mode
Lightweight integration via Claude Code hooks and CLI. No daemon required — perfect for quick setups. See the Skills guide.
SSH Tunnel Mode
Hub-initiated SSH tunnel to your server. One bootstrap command, zero inbound firewall rules.
One-Time Registration Tokens
Generate tokens with configurable expiry. Copy the install command, paste on your server, done.
Live Health Monitoring
Heartbeats with CPU, memory, and disk stats. Servers marked offline when they disconnect.
Cross-Platform
Linux (systemd), macOS (launchd), Windows (Windows Service or standalone .exe).
Permission Governance & Approval Rules
Control what your AI agents can and cannot do — with full audit trail.
Approval Rules
Define regex-based rules for which tools agents can use. Scope to the entire tenant, specific projects, or individual servers.
Allow / Deny Actions
Priority-based rule matching determines if a tool invocation is allowed or blocked.
Real-Time Enforcement
Rules are pushed to agents instantly when updated. PreToolUse hooks enforce them before every tool execution.
Permission Dashboard
See every tool permission decision across all agents. Filter by project, session, tool, and decision type.
Audit Trail
Every allowed, denied, and asked decision is logged with full context.
Project & Team Management
Organize servers, sessions, and tasks into projects. Invite your team with role-based access.
Projects
Group servers, sessions, and tasks into logical workspaces.
Team Roles
Owner (full control), member (read-write), viewer (read-only).
Email Invitations
Invite team members by email. They join automatically on signup.
Project Workbench
Drag-and-drop sortable terminal tiles showing all active sessions in a project.
File Transfer
Move files between your browser and remote agent sessions.
Upload to Session
Drag-and-drop files from your browser into a running session (up to 10MB).
Download from Session
Agents can push files to your browser for download.
Integrity Verification
SHA-256 checksums on all transfers.
CLI Support
Transfer files from the command line between sessions.
SSH Access Grants
Give one server SSH access to another — securely, with full lifecycle management.
Inter-Server SSH
Create grants from a source server to a target server, tunneled through the hub.
Privilege Levels
Standard, elevated (sudo), or root access.
Automatic Key Management
SSH key pairs generated, distributed, and revoked automatically.
Grant Lifecycle
Active grants can be revoked instantly. Expired grants clean up automatically.
Telegram Bot Integration
Get notified on your phone. Respond from Telegram.
Per-Project Bots
Each project gets its own Telegram bot for focused notifications.
Session Focus
Watch a specific session's live output directly in Telegram.
Smart Formatting
TUI noise filtering strips box-drawing and spinners. Messages are edited in-place.
Interactive Buttons
Allow/Deny permission requests directly from Telegram.
External Service API & Webhooks
Integrate Orchestratia into your existing toolchain.
Service API Keys
Generate scoped API keys per project for external systems.
Task API
Create, read, and list tasks from external services — CI/CD, scripts, or other platforms.
External Reference Tracking
Correlate Orchestratia tasks with Jira tickets, Linear issues, or any external ID.
Outbound Webhooks
Receive HMAC-SHA256-signed webhook events for task lifecycle changes with retry logic.
Voice Chat
Talk to your terminal. Browser-native, zero API cost.
Speech-to-Text
Dictate commands into terminal sessions using your browser's microphone.
Text-to-Speech
Have terminal output read aloud.
Hold-to-Talk
Push-to-talk interaction for hands-free operation.
Authentication & Security
Enterprise-grade security throughout.
Email/Password + OAuth
Sign up with email or connect Google and GitHub accounts.
CAPTCHA Protection
Cloudflare Turnstile on registration to prevent bots.
Role-Based Access Control
Project-level roles enforce who can see and do what.
TLS Everywhere
HTTPS-only with modern TLS, HTTP/2, and strict transport security.
Full Audit Trail
Every permission decision, task change, and system event is logged.
Mobile & Progressive Web App
Use Orchestratia from any device.
Installable PWA
Add to home screen on mobile or desktop for app-like experience.
Responsive Design
Every page adapts to mobile, tablet, and desktop.
Touch-Optimized Terminal
Mobile-friendly terminal controls for on-the-go monitoring.
Agent CLI Tool
A full-featured command-line tool for agents and orchestrators working inside terminal sessions.
Task Workflow
orchestratia task check, start, complete, or fail. Full lifecycle from the command line.
Plan Submission
Submit execution plans for review before coding begins.
Notes & Help
Add notes, request human intervention, respond to interventions.
File Transfer
Send files between sessions directly from the CLI.
Dependency Management
Add and remove task dependencies from the command line.
Smart ID Resolution
Type just the first 4 characters of a task ID — the CLI resolves the rest.
Agent-Agnostic
Works With the Tools You Already Use
Claude Code
Anthropic's CLI coding agent
Gemini CLI
Google's terminal AI assistant
Codex
OpenAI's coding agent
Any CLI Agent
If it runs in a terminal, we support it
Start orchestrating your AI agents today
Deploy the agent daemon, connect your servers, and take control of your AI coding fleet.