aictl

Native AI agent for your terminal and macOS desktop. Run cloud and local models through one secure runtime.

Project website: aictl.app — source in website/.

User guides (main components):

• Desktop (macOS): https://aictl.app/desktop.html
• Terminal: https://aictl.app/terminal.html
• Server: https://aictl.app/server.html

You can use aictl on your desktop and in your terminal — with a single configuration and feature parity between CLI and desktop app — or as an HTTP server with LLM proxy and security features.

Install

curl -sSf https://aictl.app/install.sh | sh

For build-from-source, Docker, optional cargo features (gguf, mlx, redaction-ner), the macOS desktop app, and uninstall instructions, see docs/INSTALL.md.

Quick start with CLI

# Interactive REPL (after running aictl --config once to set provider + key):
aictl

# Single-shot message:
aictl --message "What is Rust?"

# Autonomous mode (no tool-call confirmation prompts):
aictl --auto --message "List the largest files in this directory"

The first launch will walk you through provider, model, and API-key setup. State (config, sessions, audit log, agents, skills) lives under ~/.aictl/.

Desktop app (macOS)

A native macOS desktop app (aictl-desktop) is available as a Tauri-based frontend over the same aictl-core engine the CLI uses — same providers, tools, agents, skills, and config. Download the latest .dmg from aictl.app/desktop.html (or the GitHub releases page), open it, and drag aictl.app into /Applications. Apple Silicon and Intel builds are published per release. For build-from-source instructions see docs/INSTALL.md.

HTTP server (aictl-server)

A second binary in this workspace, aictl-server, exposes the same provider catalogue over an OpenAI-compatible HTTP endpoint with redaction, prompt-injection blocking, audit, and a master-key gate. Pure proxy — no agent loop, no tools, no agents/skills/sessions. See docs/SERVER.md for the full reference.

curl -fsSL https://aictl.app/server/install.sh | sh
aictl-server     # listens on 127.0.0.1:7878 by default; prints master key on first launch
aictl --serve    # convenience shortcut from the CLI; forwards trailing args after `--`

aictl-server can also be used as the upstream for the CLI itself (centralized provider keys + audit on the server, single master key on every client laptop), or as Claude Code's Anthropic-compatible inference provider — including opt-in cross-provider routing that translates Claude Code's POST /v1/messages to/from OpenAI, Gemini, Ollama, and the rest.

Documentation

Topic	File
Install, build, Docker, desktop app	docs/INSTALL.md
CLI flags, REPL slash commands, sessions, examples	docs/USAGE.md
Config keys, API keys, secure key storage, security knobs	docs/CONFIG.md
Supported providers and per-token pricing	docs/PROVIDERS.md
Realistic chat / coding-agent workday cost estimates	docs/LLM_PRICING.md
Built-in tools, image capabilities, security gate	docs/TOOLS.md
Agents, skills, plugins, hooks, MCP servers	docs/EXTENSIONS.md
Coding-agent mode (experimental)	docs/CODING_AGENT.md
HTTP server (aictl-server) reference	docs/SERVER.md
Architecture diagrams	docs/ARCH.md
Roadmap	ROADMAP.md
Design notes	docs/DESIGN.md

Tests

cargo test

Unit tests cover core logic across six modules: commands (slash command parsing), config (config file parsing), tools (tool-call XML parsing), ui (formatting helpers), llm (cost estimation and model matching), and security (shell validation, path validation, output sanitization).

Claude Code Skills

This project includes Claude Code skills for common workflows in .claude/skills/:

Skill	Description
/commit	Commit staged and unstaged changes with a clear commit message
/update-docs	Update README.md, CLAUDE.md, and the docs/ tree to match the current project state
/evaluate-rust-quality	Audit code quality, idiomatic Rust usage, and best practices
/evaluate-rust-security	Audit security posture, injection risks, and credential handling
/evaluate-rust-performance	Audit performance patterns, allocations, and CLI responsiveness
/project-stats-report	Generate a project statistics report (LOC, commit activity, contributors, etc.)
/sync-models	Check each provider for newly released models and update the supported set and README
/create-hook	Add a lifecycle hook to ~/.aictl/hooks.json (event, matcher, command, timeout)
/add-mcp-server	Connect an MCP server by adding an entry to ~/.aictl/mcp.json

Evaluation reports are saved to .claude/reports/ with timestamped filenames.

Support & consulting

aictl is free to use for non-commercial purposes (see License). The CLI, desktop app, and HTTP server are open source and developed in the open — no paid tiers, no feature gating, no telemetry.

If you or your team need help with any of the following, get in touch at piotr@wittchen.io:

• Hosted aictl-server — managed multi-user proxy with shared keys, audit logs, and per-user budgets, deployed on your infra or mine.
• On-prem deployment & setup — installing, configuring, and integrating aictl into your stack (corporate proxies, SSO, internal MCP servers, custom redaction rules).
• Custom features & integrations — provider adapters, MCP servers, plugins, redaction detectors, or domain-specific agents and skills tailored to your workflow.
• Custom AI Agents development — design and build bespoke AI agents (tools, prompts, evaluations, deployment) tailored to your domain, whether on top of aictl or as standalone systems.
• Commercial licensing — a license for use inside a for-profit organization (the default PolyForm Noncommercial License does not cover this).
• Support contracts — SLA-backed support, security reviews, and prioritized bug fixes.

License

This project is licensed under the PolyForm Noncommercial License 1.0.0. It is free to use for non-commercial purposes, including personal use, research, education, and use by non-profit organizations. For commercial use, please contact piotr@wittchen.io.