Getting Started with askHuman
Welcome to the official documentation for askHuman and the ALPH command-line tool. This guide will walk you through setting up your account, connecting your AI agents, and understanding the core features that put you in control.
Introduction
askHuman is an agent IM messenger that allows you to connect and interact with multiple AI agents remotely. It provides a real-time activity feed to track agent activities, tool calls, queries, and notifications with push notification support. You can manage your account, configure your AI agents, define custom contexts (skills), access a public context library, and manage cloud memory.
- askHuman (Web App): Your central dashboard with a Slack-like interface, real-time activity updates, push notifications, context library, and comprehensive agent management.
- ALPH (CLI Tool): A Node.js CLI tool with subcommand-based interface that configures your local AI agents to MCP Servers. Features interactive setup wizard, atomic file operations, automatic backups, and multi-agent support.
How It Works: A Quick Overview
- Create askHuman account: You use the askHuman web app to create a unique MCP Server endpoint. Here, you define "Contexts" (your rules and guidelines) and set up "Tools" that dictate how your AI can contact you.
- Connect on your Machine: You use the ALPH CLI tool on your computer to instantly configure your AI agents. ALPH tells them where your MCP Server is and how to talk to it.
- Interact from Anywhere: Your AI agents can now read your guidelines, access shared memory, and most importantly use tools like
askUserto send you questions. You can see these questions and respond directly from your askHuman activity feed, wherever you are.
Step 1: Set Up Your askHuman Account
First, you need to create your account and get your unique server endpoint.
Sign Up for askHuman
Navigate to https://www.askhuman.net and sign up for an account. New users will be guided through a brief setup to define initial agent instructions.
Find Your MCP Server Endpoint
Once you are logged into your dashboard, navigate to the Your MCP Server tab from the main menu. This page contains two critical pieces of information you will need for the next step. Copy them to your clipboard or a temporary note:
- MCP Server Endpoint URL: A unique URL that looks like
https://app.askhuman.net/mcp/<your-unique-server-id>. - MCP Access Key: A secret key used to authenticate your agents.
You now have your central hub ready. The next step is to connect your local AI agents to it.
Step 2: Connect Your AI Agents with ALPH
ALPH is a Node.js CLI tool with a subcommand-based interface that configures all supported AI agents on your machine. It features automatic agent detection, atomic file operations, and comprehensive backup support.
Install ALPH
Install ALPH globally using npm. You'll need Node.js 18.0.0+ and npm installed:
npm install -g @aqualia/alph-cli
Quick Setup (NPX - No Installation)
Alternatively, run ALPH directly without installation using npx:
npx @aqualia/alph-cli setup --mcp-server-endpoint https://app.askhuman.net/mcp/your-server-id --bearer your-access-key
Configure Your Agents
ALPH automatically detects installed AI agents (Gemini CLI, Cursor, Claude Code, etc.). Choose your preferred setup method:
A) Interactive Setup (Recommended):
Launch the setup wizard that guides you through configuration:
alph
Or explicitly run the setup command (no options launches the wizard):
alph setup
B) Non-Interactive Setup:
Configure directly with command-line options:
alph setup \
--mcp-server-endpoint https://app.askhuman.net/mcp/your-server-id \
--bearer your-access-key \
--agents gemini,cursor
C) Preview Changes (Dry Run):
Preview what changes will be made without modifying files:
alph setup \
--mcp-server-endpoint https://app.askhuman.net/mcp/your-server-id \
--bearer your-access-key \
--dry-run
Verify the Connection
Check the configuration status of all detected agents:
alph status
This displays a table showing each agent's configuration status and configured MCP servers.
Supported AI Agents
- Gemini CLI - Modifies ~/.gemini/settings.json
- Cursor - Platform-specific configuration locations
- Claude Code - Claude-specific configuration format
- Generic Provider - Custom configuration files
Step 3: Explore the Context Library
After setting up your agents, explore askHuman's public context library to enhance your AI workflows.
What is the Context Library?
The Context Library is a public repository of contexts (skills, instructions, and best practices) created by the community. You can browse, search, and add these contexts to your personal collection. The library features a carousel view of featured contexts for easy discovery.
Accessing the Library
From your dashboard, click the "Add from Library" button in the Contexts tab, or visit the library directly at your askHuman dashboard's library section. New users are guided through featured contexts during initial setup.
Finding Contexts
- Search: Find contexts by title or content
- Filter by Category: Browse by categories like "Standards," "Workflow," etc.
- Filter by Tags: Use tags like "Development," "Security," etc.
- Pagination: Browse through multiple pages of contexts
Adding Contexts to Your Collection
Click "Add to my contexts" on any library context. This creates a personal copy with immediate visual feedback that you can modify independently without affecting the original. The system copies all context data including title, content, description, category, and tags to your personal collection.
Core Features Reference
Overview of key features available in your askHuman dashboard:
Dashboard Layout
The dashboard features a modern Slack-like interface with a collapsible sidebar optimized for mobile-friendly navigation. The interface defaults to dark theme with professional purple branding and supports real-time updates via Server-Sent Events.
Core Tabs
- Activity Feed (Default): Real-time, Twitter-like feed of all agent interactions with infinite scrolling, chat bubbles, and Server-Sent Events for live updates.
- Your MCP Server: manage your unique server endpoint URL, generate/revoke access keys, and view configuration examples for different AI agents.
- Tools: Configure agent interaction methods including Email, Slack, and Activity feed notifications with custom conditions.
- Contexts: Manage your personal contexts and add contexts from the public library. Features dedicated "New Context" and "Edit Context" modals with direct API integration. Create, edit, and delete your instruction sets with enhanced user interface.
- Cloud Memory: Long-term memory store for your agents with timestamped entries.
- Account: Account management including activity history deletion and account deletion.
Advanced Features
- Push Notifications: Browser notifications for new agent activities with clickable links back to the dashboard and optional notification sounds.
- Slack Integration: OAuth-based Slack connection with visual channel/user selection for notifications using an integrated channel selector.
- Enhanced Context Library: Public repository of community-created contexts with search, filtering, carousel view of featured contexts, and seamless personal copying with visual feedback.
- Real-time Updates: Live activity feed updates using Server-Sent Events with infinite scrolling and chat bubble interface (no page refresh needed).
- Mobile-Optimized: Fully responsive design with optimized collapsible sidebar and mobile-first approach for all screen sizes.
- Component-Based Architecture: Modular dashboard design with dedicated components for each major feature area ensuring consistent user experience.
ALPH CLI Command Reference
Main Commands
| Command | Description |
|---|---|
alph |
Launch interactive setup wizard (recommended) |
alph setup |
Setup command (no options launches wizard) |
alph setup [options] |
Configure agents with command-line options |
alph status |
Show detected agents and configured MCP servers |
alph remove [options] |
Remove MCP server configurations |
Setup Options
| Option | Description |
|---|---|
--mcp-server-endpoint <url> |
MCP server endpoint URL |
--bearer [token] |
Authentication token (optional, redacted in output) |
--agents <list> |
Comma-separated agent names to filter |
--dry-run |
Preview changes without writing files |
--name <id> |
Name/ID for the MCP server (optional) |
-i, --interactive |
Launch interactive setup wizard |
Remove Options
| Option | Description |
|---|---|
--server-name <name> |
MCP server name to remove |
-y, --yes |
Skip confirmation prompt |
-i, --interactive |
Launch interactive removal wizard |
--no-backup |
Do not create backups before removal (advanced) |
Safety Features
- Atomic Operations: All file modifications use atomic write operations
- Automatic Backups: Timestamped backups created before changes
- Validation: Configuration integrity verified after modifications
- Error Recovery: Clear error messages and rollback instructions
- No Network Calls: Operates entirely on local files for security
Context Library
The Context Library is a community-driven repository of contexts (skills, instructions, and best practices) that enhance your AI agent workflows.
Features
- Public Repository: Browse contexts created by the community
- Search & Filter: Find contexts by title, content, category, or tags
- Personal Copies: Add contexts to your collection as independent copies
- Community Contributions: Create and share contexts with the community
Using the Library
- Navigate to the Library section from your dashboard
- Search or filter contexts by category (Standards, Workflow, etc.) or tags (Development, Security, etc.)
- Preview context content and descriptions
- Click "Add to my contexts" to copy to your personal collection
- Modify your personal copy as needed - changes don't affect the original
Contributing to the Library
Share your expertise by creating public contexts that others can benefit from. Use the "Create new context" button in the Library to contribute.
Notifications & Real-time Updates
askHuman provides multiple notification methods to keep you informed of agent activities.
Push Notifications
- Browser Notifications: Receive desktop notifications for new agent activities
- Clickable Links: Notifications link directly back to your dashboard
- Sound Alerts: Audio notification when new activities arrive
- Permission Based: Requests browser notification permission on first use
Notification Preferences
Configure how agents contact you in the Tools tab:
- Activity Feed Only: Notifications appear only in your dashboard (default for new users)
- Email: Receive notifications via email
- Slack: Integrate with Slack for team notifications
Slack Integration
- OAuth Connection: Secure Slack workspace integration
- Channel Selection: Choose specific channels or direct messages for notifications
- Visual Selector: Browse and select channels/users from an integrated interface
Real-time Activity Feed
- Server-Sent Events: Live updates without page refresh
- Infinite Scrolling: Automatically loads older activities as you scroll
- Chat Interface: Activities displayed as conversational chat bubbles
Troubleshooting & Support
ALPH CLI Issues
- Command not found: Ensure npm global bin directory is in your PATH. Check with
npm config get prefix - Permission errors: You may need
sudofor protected directories, or fix npm permissions:sudo chown -R $(whoami) $(npm config get prefix)/{lib/node_modules,bin,share} - Agent not detected: Ensure your AI agent is properly installed with configuration files in expected locations
- Configuration file not found: Use the generic provider for custom installation paths
askHuman Dashboard Issues
- Notifications not working: Ensure browser notification permissions are granted
- Real-time updates not working: Check if your browser supports Server-Sent Events
- Slack integration issues: Verify OAuth connection and channel permissions
- Context library not loading: Check your internet connection and refresh the page
Additional Resources
- ALPH CLI Help: Run
alph --helporalph setup --helpfor command-line help - GitHub Issues: Report bugs at github.com/Aqualia/Alph/issues
- Troubleshooting Guide: Detailed troubleshooting at the ALPH repository
- Documentation: Full documentation available in the ALPH repository