crestalnetwork/intentkit
Intentkit
IntentKit is an open-source platform for deploying and managing a collaborative team of AI agents that can work together to complete complex tasks. It provides a self-hosted agent orchestrator that coordinates multiple agents through a modular pipeline of entrypoints, orchestration, and storage, all running as containerized services using Docker Compose or Swarm for production-grade deployment.
The platform distinguishes itself by offering a plugin-based system for extending agent capabilities without modifying the core codebase, along with built-in integrations for connecting agents to social platforms like Discord, Twitter, and Telegram. It also supports Web3 blockchain operations, enabling agents to perform on-chain transactions, and provides a scheduling system for running autonomous agent tasks on a recurring timetable without manual intervention.
The system includes a Discord AI bot framework with conversation memory, rate limiting, and channel management, and exposes agent APIs for external integration. It persists agent state and memory in PostgreSQL with in-memory caching, and offers observability through execution trace export for debugging and performance analysis. The platform can be deployed as a local preview for evaluation or scaled to a multi-node cluster using Docker Swarm.
Features
- Multi-Agent Orchestration Platforms - Orchestrates a collaborative team of AI agents through a modular pipeline with plugin support.
- Collaborative Agent Clusters - Deploys a self-hosted cluster of collaborative AI agents that work together to complete complex tasks.
- Autonomous Task Scheduling - Schedules autonomous AI agents to execute tasks on a recurring timetable without manual intervention.
- Agent Integration APIs - Provides a dedicated API for each agent so external systems can invoke its capabilities.
- Agent Integrations - Connects agents to Twitter and Telegram so they can interact with users through those channels.
- Agent Memory Persistence - Persists agent state, credentials, and conversation memory in PostgreSQL with in-memory caching.
- Agent Capability Extensions - Offers a plugin-based system for extending agent capabilities without modifying the core codebase.
- Cluster Configurations - Provides environment-driven configuration for deploying a collaborative team of AI agents.
- Agent Skill Frameworks - Provides a pluggable skill system for adding new abilities to agents to perform custom tasks.
- Agent Configurations - Provides structured configuration files and settings to define agent behavior, parameters, and providers.
- AI Agent Orchestrators - Coordinates a team of collaborative AI agents that can call each other and share context to complete complex tasks.
- Agentic Tool Orchestration - Loads isolated tools dynamically based on agent configuration and orchestrates language model interactions with prompt engineering.
- Social Platform Adapters - Ships built-in adapters for connecting agents to Discord, Twitter, and Telegram for user interaction.
- External Service Integrations - Connects agents to platforms like Twitter and Telegram via adapters that handle input/output transformations and rate limiting.
- Multi-Agent Orchestration Layers - Routes incoming requests through separate entrypoint, orchestration, and storage layers for flexible multi-agent coordination.
- Multi-Agent Orchestrators - Manages a collaborative team of AI agents through a modular pipeline that separates entrypoints, orchestration, and storage.
- Social Platform Integrations - Connects the agent cluster to blockchain networks and social media services for automated interactions.
- Environment Variable Configuration - Reads all agent cluster settings from environment variables or AWS Secrets Manager at startup without hardcoded configuration files.
- Agent Cluster Deployments - Deploys the full AI agent stack as containerized services using Docker Compose or Swarm.
- Docker Swarm Orchestration - Launches the full agent stack as containerized services using Docker Compose or multi-node Swarm orchestration.
- Agent Run Scheduling - Executes agent tasks automatically on a recurring timetable without manual intervention.
- Agent Orchestrators - Coordinates multiple AI agents through a modular pipeline of entrypoints, orchestration, and storage.
- Self-Hosted Deployments - Deploys the full AI agent cluster on user-owned infrastructure using Docker Compose or Swarm.
- Discord AI - Provides a complete framework for connecting AI agents to Discord as bots with memory and rate limiting.
- Discord Integrations - Connects an AI agent to Discord as a bot that responds to messages in servers and DMs through a persistent WebSocket connection.
- Configuration-Driven Plugin Loaders - Loads isolated tools dynamically from configuration at runtime without modifying the core codebase.
- Blockchain AI Agent Integrations - Enables AI agents to perform on-chain blockchain transactions and Web3 operations.
- Conversational Bot Development - Handles mentions, replies, typing indicators, message splitting, and per-user or per-channel memory for Discord interactions.
- Agent Execution Traces - Sends detailed execution traces of agent runs to an external observability platform for debugging.
- Persistent Database-Backed Caches - Stores agent configurations, memory, and conversation state in PostgreSQL while caching frequently accessed data in memory.
- Agent Libraries - Imports the agent cluster functionality into existing Python projects to add collaborative AI capabilities programmatically.
- Docker Compose Deployments - Launches the API server, workers, and supporting services using Docker Compose for container orchestration.
- Agent Cluster Launchers - Launches the API server using Docker Compose or Swarm to begin managing a team of AI agents.
- Discord Gateway Clients - Connects agents to Discord and other platforms through persistent WebSocket connections for real-time message handling.
- Bot Access Controls - Restricts bot activity to specific guilds, channels, or users with special owner handling.
- Blockchain Transaction Execution - Performs Web3 transactions and blockchain interactions through optional crypto integrations.
- Agent Skill Extenders - Extends agent capabilities through a pluggable skill system without modifying the core codebase.
- Platform-Specific Rate Limiters - Respects Discord's global, per-channel, and gateway rate limits automatically to prevent bot throttling.
- Trace Exporters - Sends detailed execution traces of agent runs to external platforms for debugging and performance analysis.
