Quick Start
Deploy your first AI Rail in under 5 minutes with our step-by-step guide.
API Reference
Complete REST API documentation with examples for every endpoint and SDK method.
Tutorials
Hands-on guides for common use cases: sentiment analysis, data extraction, content generation, and more.
Core Concepts
Understanding these foundational concepts will help you make the most of RailMind.
Rails
A Rail is a named, versioned AI pipeline with a defined input/output schema, model binding, and guardrail configuration. It is the fundamental unit of deployment in RailMind.
Schemas
Schemas define the structure of your AI's output. RailMind validates every LLM response against your schema before it reaches downstream systems, ensuring type safety and structural integrity.
Guardrails
Guardrails are safety policies attached to a Rail. They include PII detection, prompt injection filtering, token limits, content moderation, and custom validation rules.
Routes
Routes define how tasks are distributed across different LLM providers. The intelligent router evaluates complexity, cost, and latency to pick the optimal model for each request.
SDK Quick Reference
RailMind provides official SDKs for TypeScript and Python.
Changelog
Recent updates and improvements to the RailMind platform.
Added multi-model cascading for intelligent routing. New fallback policies and cost-aware model selection. Dashboard now supports dark/light themes.
Introduced autonomous RAG healing module. Stale vector index detection now runs on configurable schedules. Python SDK reaches feature parity with TypeScript.
OpenTelemetry integration for full distributed tracing. New guardrail plugins for HIPAA and SOC 2 compliance. Performance improvements across the board.