Your secure personal AI assistant
IronClaw is built on a simple principle: your AI assistant should work for you, not against you. With defense-in-depth security, local-first data storage, and self-expanding capabilities, IronClaw is the AI assistant you can actually trust.
WASM sandbox active
Prompt injection defense enabled
Credentials protected
Local-first storage
Get started in minutes
Follow these steps to install and configure IronClaw
Install IronClaw
Install IronClaw using your preferred package manager or download the binary directly:
Set up your database
IronClaw requires PostgreSQL with the pgvector extension for persistent storage:
You can also use libSQL/Turso as an embedded database alternative. See the database configuration guide for details.
Run the onboarding wizard
Configure IronClaw with the interactive setup wizard:The wizard will guide you through:
- Database connection setup
- LLM provider authentication (NEAR AI by default)
- Secrets encryption configuration
- Optional channel setup (Telegram, HTTP, etc.)
Security-first architecture
Multiple layers of defense protect your data and credentials
WASM sandbox
Untrusted tools run in isolated WebAssembly containers with capability-based permissions and resource limits
Prompt injection defense
Pattern detection, content sanitization, and policy enforcement prevent malicious prompt injection attacks
Credential protection
Secrets are encrypted with AES-256-GCM and never exposed to tools or LLMs—injected only at the host boundary
Data protection
All data stored locally in your PostgreSQL database with no telemetry, analytics, or data sharing
Powerful features
Everything you need for a personal AI assistant
Multi-channel support
Connect via REPL, HTTP webhooks, Telegram, web gateway, or Signal—all channels work simultaneously
Dynamic tool building
Describe what you need and IronClaw builds it as a WASM tool with automatic capability detection
Persistent memory
Hybrid search with full-text and vector embeddings for context-aware conversations
Automated routines
Schedule background tasks with cron expressions or trigger actions from webhooks and events
Docker sandbox
Execute complex jobs in isolated containers with per-job tokens and orchestrator/worker pattern
MCP integration
Connect to Model Context Protocol servers for additional capabilities like GitHub, Gmail, and more
Explore the documentation
Find guides, references, and examples for every feature
Core concepts
Learn about IronClaw’s philosophy, architecture, and security model
Configuration
Configure LLM providers, database connections, and environment variables
CLI reference
Complete reference for all commands, subcommands, and flags
API reference
Rust API documentation for embedding IronClaw in your own applications
Ready to get started?
Install IronClaw and start building with a secure AI assistant that protects your data
Get started now