AetherShell Documentation
The world's first multi-agent shell with typed functional pipelines
Welcome to AetherShell
AetherShell is a next-generation shell that combines:
- 🤖 Multi-Agent Orchestration - Deploy AI agent swarms
- 💎 Typed Functional Pipelines - Hindley-Milner type inference
- 🎨 Multi-Modal AI - Images, audio, video processing
- 💬 AgenticBinary Protocol - Maximum information density
- 🔧 MCP Integration - Standardized tool access
What Makes AetherShell Unique?
🤖 Multi-Agent Orchestration
The ONLY shell that can deploy swarms of AI agents with different models and capabilities:
swarm({
agents: [
{id: "analyzer", model: "gpt-4o", role: "analyze"},
{id: "writer", model: "claude-3", role: "write"},
{id: "reviewer", model: "llama3", role: "review"}
],
strategy: "router",
max_steps: 10
})
💬 AgenticBinary Protocol
Maximum information density binary protocol with 16 semantic opcodes:
// Encode agent message (3-5x compression)
msg = ab_encode("command", "delegate", "task:analyze_data")
// Decode with full details
decoded = ab_decode(msg)
print(decoded.opcode) // "DELEGATE"
💎 Typed Functional Pipelines
Structured data flows with type inference:
// Type-safe data transformations
[1,2,3,4,5]
| map(fn(x) => x * 2)
| where(fn(x) => x > 5)
| reduce(fn(a,b) => a + b, 0)
// Result: 24 (type: Int)
Getting Started
1
Install AetherShell
cargo build --release
./target/release/ae
2
Run Your First Command
print("Hello, AetherShell!")
[1,2,3] | map(fn(x) => x * 2)
3
Explore AI Features
agent("List files in current directory", "ls", 5)