Built on GitHub Copilot SDK

Your Copilot CLI,
visualized.

A visual control center for the GitHub Copilot CLI. Manage parallel sessions, agents, and workflows from one window — with cross-session search and mobile access wherever you are.

Get started → View on GitHub

⚡ Every Copilot CLI session, auto-discovered. Continue from CLI, Desktop, or Mobile.

🖥️ Core Experience

Multi-session visual management

Run multiple Copilot CLI sessions simultaneously in a tabbed interface. Each session gets its own system prompt, model, tools, MCP servers, and working directory. Monitor all agents at a glance — no more juggling terminals.

  • ▸ Per-session context isolation — system prompt, model, tools, MCPs, CWD
  • ▸ Unread indicators across parallel sessions
  • ▸ Switch models, enqueue messages, attach files & paste images, set modes
  • ▸ Rich rendering — Markdown, syntax highlighting, Mermaid diagrams, reasoning steps
Multi-session management with agent library and session tabs
Push notification on lock screen
Session list on mobile
Chat conversation on mobile
📱 Every CLI session, on your phone

Your CLI sessions, in your pocket

Every Copilot CLI session — Console-launched and native terminal — shows up on your phone the moment it starts. Push notifications when any agent finishes or asks a question. Reply from the chat view. Installable as a PWA.

$ copilot-console --expose --no-sleep
# Scan QR from Settings, done.

⚡ Mobile session access and notifications work for every session — terminal or Console. Tap a notification, continues from where the agent left off.

Slash command palette with /fleet, /compact, /agent
Agent picker via /agent slash command
Inline /fleet chip in input box
⚡ Type / to Explore

Slash commands

Type / to open the command palette. Pick a command, get an inline chip with contextual hints. No flags to memorize, no syntax to look up.

  • ▸ /fleet — fire parallel sub-agents, watch multiple AI streams simultaneously
  • ▸ /compact — compress session context to free up the conversation window
  • ▸ /agent — set the primary agent for your session, enabling spec-driven frameworks like SpekKit and GSD that rely on a selected agent to orchestrate workflows
💬 Interactive Q&A

Agents ask you when they need to

When an agent needs clarification, it asks a structured question right in the conversation. Works seamlessly on desktop and mobile.

  • ▸ Structured elicitation — agents ask targeted questions mid-conversation
  • ▸ Works on desktop & mobile — answer from anywhere
  • ▸ Resilient — pending questions survive connection drops on mobile
Agent asking the user a structured question mid-conversation
📂 Organized by Project

Project facilitation

Sessions automatically group by project folder. Search across all conversations with keyword highlighting. Pin important messages for quick reference. Your AI history, organized automatically.

Project-based session organization
Cross-session search with keyword highlighting
Pinned messages panel
🔀 Multi-agent pipelines

Workflows

Chain agents into deterministic pipelines with YAML — built on Microsoft Agent Framework. See the workflow as a Mermaid diagram. Branch with If/Switch, loop with Foreach, pause for human input. Run history is replayable; failures surface as sticky toasts. Optional Power Fx =expressions for richer logic.

Workflow run with Mermaid diagram and event stream

And much more

Everything you need to orchestrate AI agents at scale.

⏰

Automations

Schedule agent runs on cron. Runs dashboard for all executions — continue any run's conversation or watch a live agent in real-time.

🤖

Agent Library

Reusable agent configs with custom prompts, models, tools, and MCP servers. Launch sessions from any agent with one click.

🤝

Agent Teams

Multi-agent delegation — compose agents into teams. Each sub-agent runs in its own context with specialized tools.

📎

Files & Images

Drag-and-drop files, paste images, attach context to any message. Give your agent the visual and textual context it needs.

🔌

MCP & Custom Tools

Manage MCP servers right from Settings. OAuth support with auto-recovery on token expiry. Drop Python functions as custom tools.

🌐

Web Browsing

Autonomous web navigation via bundled Playwright MCP. Fill forms, extract data, take screenshots — all through natural language.

🔔

Desktop Notifications

Know when agents finish. Native desktop notifications so you never miss a completed task, even when Console is in the background.

📂

Open With

Open project folders in VS Code, Terminal, or Explorer with one click. Jump straight from Console into your favorite editor.

👁️

Full Visibility

See what your agents are doing — reasoning steps, live status banners, sticky error alerts, and replayable run history. No surprises, no black boxes.

One command to install

Works on Windows, macOS/Linux—checks prerequisites, installs dependencies, and you’re all set. 

# Install (or upgrade)
irm https://raw.githubusercontent.com/sanchar10/copilot-console/main/scripts/install.ps1 | iex

# Start
copilot-console

# With mobile companion
copilot-console --expose --no-sleep
View on GitHub → Manual install guide