agent-air is an open-source SDK and Runtime for building AI agents running in a Terminal UI or on the server. agent-air handles everything you need: LLM communication, multiple sessions, markdown rendering, widgets & layouts, tools, skills, MCP, permissions and much more so you can focus on building your agent.
Add agent-air to your Cargo.toml:
[dependencies]
agent-air = "0.7.0"Build a working AI agent in just a few lines of code:
use agent_air::agent::AgentAir;
use agent_air::tui::AgentAirExt;
const SYSTEM_PROMPT: &str = "You are a helpful AI assistant.";
fn main() -> std::io::Result<()> {
AgentAir::with_config("my-agent", "~/.config/my-agent", SYSTEM_PROMPT)?
.into_tui()
.run()
}
- High Performance Rust Core - Async-first architecture powered by Tokio with zero-cost abstractions
- Multi Model Architecture - 75+ providers and 500+ models with runtime switching
- Modular Architecture - Terminal UI, Desktop, or Web-based frontends with event-driven API
- Advanced Session Management - Isolated concurrent sessions with independent history and tool registrations
- Advanced Context Management - Threshold-based, on-demand, or LLM-driven compaction strategies
- Flexible Tool Management - Custom tools with JSON schema and parallel batch execution
- Powerful Permissions Framework - Target + Level model with batch requests and recursive path grants
- Streaming & Markdown Rendering - Real-time SSE parsing with CommonMark-compliant rendering
- Widget System - Pre-built widgets: StatusBar, ChatView, TextInput, QuestionPanel, PermissionPanel, SessionPicker
- Slash Command Framework - Built-in and custom commands with interactive popup and real-time filtering
- Agent Skills Support - Modular skill system with reusable, composable behaviors
- MCP Protocol - Model Context Protocol support for external tool integrations
- Custom Keybindings - Standard and Emacs bindings with full KeyHandler trait for custom bindings
- Error Recovery - Graceful degradation with retries and actionable error messages
Contributions in any form (suggestions, bug reports, pull requests, and feedback) are welcome. If you've found a bug, you can submit an issue or email me at rsingh@arrsingh.com.
This project is dual-licensed under the MIT LICENSE or the Apache-2 LICENSE:
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.
Contact: rsingh@arrsingh.com