pastureprotocol.com
Completely free to install and use
Your private AI companion runtime
Pasture Protocol installs a local agent system on your computer: chat from WhatsApp, Telegram, or the dashboard; choose local or cloud models; give each agent the tools it should actually be allowed to use.
Install in one line
curl -fsSL https://raw.githubusercontent.com/bishwashere/pastureprotocol/master/install.sh | bashWindows (PowerShell)
iwr -useb https://raw.githubusercontent.com/bishwashere/pastureprotocol/master/install.ps1 | iexThen run pasture start and pasture dashboard
Demo screens
A local dashboard for the whole runtime
These are the screenshots you added: projects, missions, tasks, reports, agents, and live dashboard views from the Pasture interface.

Mission Control overview
Track projects, missions, labeled tasks, agent progress, and recent movement.

Mission Control board
See needs attention, completed work, active research, and recent movement in one view.

Tasks
Review mission tasks by status, assignee, labels, progress, and completion window.

Agents
See each agent, last task, active state, and how much work each teammate has completed.

Projects
Track project setup, connectors, branches, updates, and long-running product context.

Missions
Open persistent missions, inspect progress, and manage tasks linked to the mission tree.

Reports
Review agent throughput, delegation counts, average execution time, and skill usage.
What Pasture actually runs
Local runtime
The bot, dashboard, workspace, auth files, reminders, and memory live under your own machine, with no Pasture cloud relay.
WhatsApp, Telegram, dashboard
Baileys powers WhatsApp, a Telegram bot handles Telegram, and the local dashboard gives you chat plus Mission Control.
Prompt-driven agent team
Agents are named personas with Markdown identity, selected skills, routing rules, and internal delegation through agent-send.
Skills that do real work
Search, browse, vision, memory, cron, GitHub, Google, Home Assistant, file tools, SSH inspection, and more are enabled per agent.
Start with the runtime model
The docs explain the real Pasture Protocol architecture: prompt files are the brain, JavaScript is the local router and executor, and agents use bounded skills to act across chat, files, memory, browser automation, and integrations.
How requests are handled →