sbhooley/ainativelang

AINL helps turn AI from "a smart conversation" into "a structured worker." It is designed for teams building AI workflows that need multiple steps, state and memory, tool use, repeatable execution, validation and control, and lower dependence on long prompt loops. AINL is a compact, graph-canonical, AI-native programming system for (READ: README)

What it solves

AINL addresses the inefficiency and instability of "prompt-loop" orchestration, where AI agents repeatedly prompt an LLM to decide the next step in a workflow. This approach often leads to broken code, high token costs for recurring jobs, and a lack of of deterministic execution traces for compliance audits (e.g., SOC 2/HIPAA).

How it works

AINL is a compact, graph-canonical AI workflow language. It allows developers (or AI coding agents) to author workflows in .ainl files, which are then compiled into a deterministic Intermediate Representation (IR) graph. The runtime executes this graph deterministically, meaning the LLM is only used for the initial authoring and specific judgment gates, rather than for every routing decision in every run.

Who it’s for

  • AI Agent Developers: Those using tools like Cursor, Claude Code, or OpenClaw who want their agents to author stable, validated workflows.
  • Enterprise Teams: Organizations requiring strict compile-time validation and tamper-evident execution traces for compliance.
  • Automation Engineers: Users running numerous recurring monitors or scheduled jobs where reducing orchestration token spend is a priority.
  • Cross-Platform Deployers: Those needing a single workflow source that can be emitted to multiple targets like LangGraph, Temporal, or FastAPI.

Highlights

  • Token Efficiency: Reduces orchestration tokens by 90-95% for recurring jobs compared to prompt-loop baselines.
  • Deterministic Runtime: Compiles once and runs many times without re-spending tokens on routing.
  • MCP Integration: Provides an MCP server allowing any MCP-compatible agent to author, validate, and run AINL workflows.
  • Multi-Target Emission: Ability to emit a single workflow source to different orchestration frameworks.
  • Agent-to-Agent (A2A) Adapter: An opt-in network adapter for agent communication.
  • HTTP Machine Payments: Support for HTTP-402 payment profiles for agentic commerce.

Related

  • Project
    gethopp/figma-mcp-bridge

    A plugin and MCP server that streams live Figma document data to AI tools, bypassing official API rate limits to enable seamless design-to-code workflows.

  • Project
    docker/model-runner

    A tool for managing and deploying AI models via Docker, streamlining the process of pulling and serving LLMs using backends like llama.cpp and vLLM.

  • Project
    sdv-dev/Copulas

    A Python library for modeling multivariate distributions and generating synthetic numerical data using copula functions.

  • Project
    Project-Platypus/Platypus

    A Python framework for evolutionary computing focused on multiobjective evolutionary algorithms (MOEAs) to solve complex optimization problems with multiple goals.