Version 0.9.0 - Edge Intelligence

The Programming Language for the AI Era

Simplex combines Erlang's fault-tolerance, Rust's memory safety, and first-class AI integration to build distributed, resilient applications. Now with Edge Hive for local-first AI, Dual Numbers for automatic differentiation, and Self-Learning optimization. Powered by Cognitive Hive AI architecture.

Get Started Learn More

Simple, Expressive Syntax

Simplex code is clean and readable. AI operations are first-class citizens, not external API calls. Actors provide natural isolation and fault tolerance.

  • Static typing with aggressive type inference
  • Ownership-based memory management (no GC)
  • Pattern matching and algebraic data types
  • Built-in support for AI operations
chatbot.sx 
actor Chatbot {
    var history: List<Message> = []

    receive Chat(user_msg: String) -> String {
        history.push(Message::user(user_msg))

        let response = await ai::complete(
            build_prompt(history)
        )

        history.push(Message::assistant(response))
        checkpoint()
        response
    }
}

Why Simplex?

Built from the ground up for the AI era, combining the best ideas from proven technologies.

AI-Native

AI operations are first-class language constructs. Complete, embed, extract, and classify with simple function calls. Automatic batching and model tiering for cost efficiency.

Distributed-First

Programs naturally decompose across VM swarms. Content-addressed code enables seamless distribution. Actors migrate transparently between nodes.

Fault-Tolerant

Embrace the "let it crash" philosophy from Erlang. Supervision trees automatically restart failed actors. Checkpointing enables resumption on any node.

Cognitive Hive AI

Orchestrate swarms of specialized Small Language Models instead of monolithic LLMs. Achieve 88-96% cost savings while maintaining quality and gaining control.

Resumable

Computations checkpoint and resume transparently. Survive spot instance terminations, network partitions, and hardware failures without losing progress.

Lightweight Syntax

Clean, readable code that compiles to efficient bytecode. Static typing with aggressive inference means fewer annotations without sacrificing safety.

Neural Gates

Learnable control flow with differentiable decision points. Train your program logic from data. Gates optimize during training, compile to zero-overhead branches.

Real-Time Learning

Specialists adapt during runtime without offline training. Learn from user feedback instantly. Federated learning coordinates knowledge across hives.

Evolving Systems

Programs that get smarter over time. Decision logic optimizes from training data. Models improve from interaction. No redeployment required.

Dual Numbers

Native automatic differentiation with zero overhead. Compute exact gradients through any function. Forward-mode AD built into the type system for machine learning.

Edge Hive

Local-first AI on any device from smartwatch to desktop. On-device inference with privacy by default. Seamless cloud fallback when needed.

Self-Learning Optimization

Hyperparameters that optimize themselves via meta-gradients. Learnable temperature schedules, adaptive learning rates, and automatic pruning thresholds.

Standing on the Shoulders of Giants

Simplex combines proven ideas from the best programming languages and systems.

Erlang/OTP

Actor model, supervision trees, "let it crash" philosophy, and battle-tested fault tolerance patterns.

Rust

Ownership-based memory management, no garbage collection, algebraic data types, and pattern matching.

Ray

Distributed computing primitives, seamless scaling across clusters, and AI/ML workload optimization.

Unison

Content-addressed code where functions are identified by hash, enabling perfect caching and distribution.

What Can You Build?

Simplex excels at building enterprise applications on a Cognitive Hive AI architecture.

Content Management (CMS)

AI-powered content creation, automatic tagging and categorization, semantic search, and intelligent content recommendations.

Customer Relations (CRM)

Intelligent lead scoring, sentiment analysis, automated response suggestions, and predictive customer insights.

Document Management (DMS)

Automatic document classification, entity extraction, summarization, and semantic document search.

Data Pipelines

Distributed data processing with fault tolerance, automatic retry, and seamless scaling across compute clusters.

AI Agents

Build autonomous agents with tool use, memory, and reasoning capabilities using specialized language model swarms.

Real-Time Processing

Event-driven architectures, stream processing, and real-time analytics with sub-millisecond actor messaging.

Dramatic Cost Savings

Cognitive Hive AI architecture delivers massive cost reductions compared to traditional approaches.

88%
Infrastructure Savings
96%
AI API Cost Reduction
$310
Monthly (vs $2,650)
10M+
Requests/Month

Reference Architecture

1000 actors, 10M messages/day: 20 spot instances + 3 coordinators + 2 GPU nodes + S3 storage. Run on commodity ARM instances (t4g.nano/micro) with 90% spot instances for maximum savings.

Ready to Get Started?

Start building AI-native, distributed applications today.

Install Simplex Read Tutorials