Junkie Documentation
Build intelligent Discord experiences with multi-agent AI orchestration, secure code execution, and comprehensive context management
Quick start
Get Junkie running in minutes with these simple steps
Configure environment variables
Create a
.env file with your Discord token and API keys:See the environment configuration guide for all available options.
Key features
Powerful capabilities for building intelligent Discord bots
Multi-agent orchestration
Deploy specialized AI agents for code execution, web search, and context analysis that work together seamlessly
E2B sandbox execution
Run Python code and shell commands in secure, isolated sandboxes with full file system access
Context-aware responses
Maintain conversation history with PostgreSQL-backed caching and intelligent context retrieval
Memory persistence
Store and recall user-specific information across sessions with automatic memory management
TL;DR summaries
Generate conversation summaries on demand with the
.tldr commandPhoenix tracing
Monitor agent performance and debug interactions with integrated observability
Architecture overview
Understand how Junkie’s components work together
Team system
Learn how the multi-agent team orchestrates specialized agents
Agent roles
Explore the specialized agents and their capabilities
Context management
Understand message caching and history retrieval
Integrations & tools
Extend Junkie with powerful external services
MCP tools
Connect to Model Context Protocol servers for extended capabilities
Web search
Access real-time information with Perplexity Sonar and Exa search
Code execution
Run code securely with E2B sandboxes and Groq Compound
User tools
Fetch Discord user profiles, avatars, and status information
Resources
Additional materials to help you succeed
API reference
Complete API documentation for all core modules
Deployment guide
Deploy Junkie to production with Docker or Railway
Ready to get started?
Follow our quickstart guide to deploy your first AI-powered Discord bot in minutes
Get Started