Open Source Ecosystem

One Protocol,
Every AI Provider.

AI-Lib is an open-source ecosystem that standardizes AI model interactions. V2 protocol with MCP, Computer Use, and multimodal capabilities — 37 providers, two runtimes, zero hardcoded logic.

Get Started View on GitHub

AI Providers

Projects

Hardcoded Logic

Runtimes

Three Projects, One Ecosystem

A specification layer defines the rules. Two runtime implementations bring them to life.

v0.7.0

AI-Protocol

The provider-agnostic specification. V2 three-ring manifests with MCP, Computer Use, and multimodal schemas — no hardcoded logic, ever.

37 provider manifests (6 V2 + 36 V1)
V2 three-layer pyramid architecture
MCP / Computer Use / Multimodal schemas
ProviderContract specification
CLI tool for validation & inspection
JSON Schema validated
Hot-reloadable configuration

Learn more

v0.8.0

ai-lib-rust

High-performance Rust runtime. ProviderDriver abstraction, MCP tool bridge, Computer Use safety, extended multimodal, and operator-based streaming pipeline.

ProviderDriver (OpenAI/Anthropic/Gemini)
MCP tool bridge with namespace isolation
Computer Use + SafetyPolicy
Extended multimodal validation
Capability Registry (feature-gated)
Operator-based streaming pipeline
185+ tests, published on Crates.io

Learn more

v0.7.0

ai-lib-python

Developer-friendly Python runtime. ProviderDriver abstraction, MCP tool bridge, Computer Use safety, extended multimodal, and full async support.

ProviderDriver (OpenAI/Anthropic/Gemini)
MCP tool bridge with namespace isolation
Computer Use + SafetyPolicy
Extended multimodal validation
Capability Registry (pip extras)
Pydantic v2 + full async/await
75+ V2 tests, published on PyPI

Learn more

Protocol-Driven Design

"All logic is operators, all configuration is protocol." Every provider behavior is declared in YAML — runtimes contain zero hardcoded provider logic.

Declarative Configuration

Provider endpoints, auth, parameter mappings, streaming decoders, and error handling — all declared in YAML manifests, validated by JSON Schema.

Operator-Based Pipeline

Streaming responses flow through composable operators: Decoder, Selector, Accumulator, FanOut, EventMapper. Each operator is protocol-driven.

Hot-Reload Ready

Update provider configurations without restarting. Protocol changes propagate automatically to runtimes. Add new providers through configuration, not code.

Ecosystem Architecture

Three layers working together — specification defines the rules, runtimes execute them, applications consume unified interfaces.

How It Works

From user request to unified streaming events — every step is protocol-driven.

Choose Your Runtime

Same protocol, different strengths. Pick the runtime that fits your stack.

Capability	AI-Protocol	Rust SDK	Python SDK
Type System	JSON Schema	Compile-time (Rust)	Runtime (Pydantic v2)
Streaming	SSE/NDJSON spec	tokio async streams	async generators
Resilience	Retry policy spec	Circuit breaker, rate limiter, backpressure	ResilientExecutor with all patterns
V2 Driver	ProviderContract spec	Box<dyn ProviderDriver>	ProviderDriver ABC
MCP	mcp.json schema	McpToolBridge	McpToolBridge
Computer Use	computer-use.json schema	ComputerAction + SafetyPolicy	ComputerAction + SafetyPolicy
Multimodal	multimodal.json schema	MultimodalCapabilities	MultimodalCapabilities
Embeddings	—	Vector operations, similarity	Vector operations, similarity
Distribution	GitHub / npm	Crates.io	PyPI
Best For	Specification & standards	Systems, performance-critical	ML, data science, prototyping