Agmente  by rebornix

iOS client for AI coding agents

Created 4 weeks ago

New!

484 stars

Top 63.7% on SourcePulse

GitHubView on GitHub
Project Summary

Agmente is an iOS client designed for interacting with coding agents that utilize the Agent Client Protocol (ACP) or Codex app-server protocols. It provides users with a mobile interface to view tool calls, results, and conversation history directly on their device, enabling on-the-go access to powerful AI coding assistants. The target audience includes developers and power users who need to monitor or interact with coding agents remotely or locally from their iOS devices.

How It Works

The application functions by establishing WebSocket Secure (wss://) connections to ACP or Codex app-server endpoints. It parses the communication protocol to display tool invocations, their outputs, and the overall conversation flow within a user-friendly transcript. For remote access, it supports configurations involving Cloudflare Tunnels and Access credentials, allowing agents running on remote hosts to be securely exposed and accessed from the iOS client.

Quick Start & Requirements

  • Primary Install/Run: Build the iOS application using Xcode. The README provides CLI commands using xcodebuild for simulators or devices.
  • Prerequisites: Xcode (latest stable recommended), macOS for building iOS applications.
  • Remote Agent Setup: Requires running npx -y @rebornix/stdio-to-ws on a remote host to expose an ACP agent (e.g., copilot --acp) via wss://.
  • Documentation: Links to setup.md and Agents.md for local setup, and docs/remote-agent.md for remote agent configuration.

Highlighted Details

  • Supports a variety of ACP agents including Copilot CLI, Gemini CLI, Claude Code adapters, Qwen, and Mistral Vibe.
  • Enables remote agent access via secure wss:// connections, with optional Cloudflare Tunnel and Access integration.
  • Provides a dedicated Swift package (ACPClient) for the ACP client logic, separating concerns.

Maintenance & Community

The provided README does not contain specific details regarding notable contributors, sponsorships, community channels (like Discord/Slack), or a public roadmap.

Licensing & Compatibility

  • License: MIT License.
  • Compatibility: The MIT license is permissive and generally compatible with commercial use and closed-source applications.

Limitations & Caveats

The project is primarily focused on iOS development, requiring macOS and Xcode for building. While it supports remote agents, the setup for exposing these agents securely might involve additional complexity with services like Cloudflare Tunnels. The README does not detail known bugs or alpha/beta status.

Health Check
Last Commit

17 hours ago

Responsiveness

Inactive

Pull Requests (30d)
20
Issues (30d)
11
Star History
487 stars in the last 28 days

Explore Similar Projects

Feedback? Help us improve.