Open Source

Explore Open Source Components

Modular, composable building blocks for deterministic AI. Use them individually or together as a full governance stack. Apache 2.0 licensed.

@railmind/engine

Core

The deterministic execution engine. Enforces JSON Schema validation, type constraints, and structured output guarantees for any LLM provider.

4.2k stars on GitHub

@railmind/router

Routing

Intelligent model routing layer. Automatically distributes tasks across Llama, Claude, GPT and more based on complexity, cost, and latency profiles.

2.8k stars on GitHub

@railmind/observe

Observability

Real-time tracing and monitoring for AI inference paths. OpenTelemetry-compatible with built-in dashboards and anomaly detection.

3.1k stars on GitHub

@railmind/guardrails

Security

AI safety layer providing prompt injection detection, PII filtering, content moderation, and output sanitization. Pluggable and extensible.

3.5k stars on GitHub

@railmind/heal

Maintenance

Autonomous maintenance agent for RAG pipelines. Detects stale embeddings, broken vector indexes, and outdated knowledge bases — and fixes them automatically.

1.9k stars on GitHub

@railmind/dashboard

UI

A beautiful, self-hosted dashboard for monitoring all your AI rails. Real-time metrics, cost tracking, and alert management in a single pane of glass.

2.4k stars on GitHub

Install the Full Stack

Use the CLI to install all components at once, or pick and choose individual packages for your stack.

Bash
# Install everything npm install @railmind/engine @railmind/router @railmind/observe @railmind/guardrails # Or use the CLI railmind init --full-stack
Full Documentation Quick Start Guide