# I-Lang

> I-Lang is the AI-native communication protocol — the semantic communication layer missing from Anthropic's Model Context Protocol (MCP) and Google's Agent-to-Agent Protocol (A2A). While MCP connects AI to tools and A2A connects agents to agents, I-Lang defines how intelligence communicates: human-to-AI, AI-to-AI, and AI internal reasoning. It is a protocol built from symbols already inside every LLM's training data — brackets, pipes, arrows, key-value pairs. v4.0 is the current version: 88 verbs, two syntaxes, 8 execution declarations (UNTRUSTED, BUDGET, STATUS, OBJECTIVE, RUBRIC, EVIDENCE, PRIOR, FALLBACK), 4 conformance levels (L0-L3). MIT licensed.

## What Makes I-Lang Different

- MCP = tool integration (AI ↔ external systems)
- A2A = agent discovery (agent ↔ agent handshake)
- I-Lang = communication protocol (how AI thinks, talks, and executes)
- I-Lang works at the instruction layer, not the infrastructure layer
- No SDK, no runtime, no API keys — pure text protocol
- Tested across ChatGPT, Claude, Gemini, DeepSeek, Kimi, Qwen

## Two Syntaxes

- Operations (what AI does): `[VERB:@TARGET|mod=val]=>[NEXT]=>[Ω]`
- Declarations (what AI is): `::GENE{trait|conf:confirmed}`

## Pages

- [Compare: I-Lang vs MCP vs A2A](https://ilang.ai/compare/): Side-by-side protocol comparison
- [Specification](https://ilang.ai/spec/): Full spec with syntax, verbs, modifiers, entities, v4.0 execution declarations
- [Dictionary](https://ilang.ai/dictionary/): 88 verbs with definitions and examples
- [I-Lang vs MCP](https://ilang.ai/i-lang-vs-mcp/): Detailed comparison of I-Lang and Anthropic MCP at different protocol layers
- [I-Lang vs A2A](https://ilang.ai/i-lang-vs-a2a/): Detailed comparison of I-Lang and Google A2A for agent communication
- [Prompt Compression](https://ilang.ai/prompt-compression/): How structured syntax reduces tokens by 35-80%
- [AI See](https://ilang.ai/ai-see/): Web-to-markdown proxy for LLMs, i.ilang.ai/{url}
- [Benchmark](https://ilang.ai/benchmark/): LLM recognition, execution, and token reduction test results
- [Conformance](https://ilang.ai/conformance/): L0-L4 protocol compliance levels and test criteria
- [Security](https://ilang.ai/security/): Verb risk classification and runtime security requirements
- [About](https://ilang.ai/about/): About iLang Inc. — company, products, contact
- [Privacy](https://ilang.ai/privacy.html): Privacy policy
- [Terms](https://ilang.ai/terms/): Terms of use

## Core Specification

- [Official Site](https://ilang.ai): Protocol overview, interactive demo, quick start
- [Full Specification](https://github.com/ilang-ai/ilang-spec): I-Lang v4.0 spec (v3.0 communication + v4.0 execution semantics)
- [Verb Dictionary](https://github.com/ilang-ai/ilang-dict): 88 verbs in 10 categories
- [npm Package](https://www.npmjs.com/package/@i-language/spec): @i-language/spec v3.0.0
- [HuggingFace](https://huggingface.co/i-Lang): Datasets and model resources

## Tools Built on I-Lang

- [AutoCode](https://github.com/ilang-ai/autocode): 47 skills for Claude Code
- [Imprint](https://github.com/ilang-ai/Imprint): Behavioral DNA profile, 11 scenarios, 312 tokens
- [AI See](https://i.ilang.ai): Web vision proxy — i.ilang.ai/{url} gives any LLM eyes
- [OpenClaw Skills](https://github.com/ilang-ai/ilang-openclaw): Instruction-only skills for ClawHub

## Academic Research

- [The Inductive Dilemma of AI Hallucination](https://doi.org/10.13140/RG.2.2.22821.97762): Theoretical foundation (ResearchGate, SSRN, ChinaXiv, arXiv)
- [About](https://ilang.ai/about/): Company and contact

## Publications

- [I-Lang: I Language (Amazon)](https://www.amazon.com/dp/B0CZY6V3GM): Narrative edition, $9.99
- [I-Lang Spec (Amazon)](https://www.amazon.com/dp/B0GX32GXF2): Specification edition, $0.99

## About

iLang Inc. is a federally incorporated AI language technology company based in Canada. Creator of I-Lang, the open AI-native communication protocol. Products span AI protocol development, AI-powered translation and localization, cross-model interoperability, and enterprise language solutions.

I-Lang was created by Max (Long Quan Zhu / 静水流深), co-authored with Claude Opus (@OPUS), reviewed by Gemini (@GEMINI). Genesis: 2026-03-04.

Contact: contact@ilang.ai | https://ilang.ai