aru

An intelligent coding assistant for the terminal, powered by LLMs and Agno agents.

0329(3)

Highlights

Multi-Agent Architecture — Specialized agents for planning, execution, and conversation
Interactive CLI — Streaming responses, multi-line paste, session management
16 Integrated Tools — File operations, code search, shell, web search, task delegation
Task Planning — Break down complex tasks into steps with automatic execution
Multi-Provider — Anthropic, OpenAI, Ollama, Groq, OpenRouter, DeepSeek, and others via custom configuration
Custom Commands and Skills — Extend aru via the .agents/ directory
MCP Support — Integration with Model Context Protocol servers

Quick Start

1. Install

pip install aru-code

Requirements: Python 3.13+

2. Configure the API Key

Aru uses Claude Sonnet 4.6 from Anthropic as the default model. You need an Anthropic API key to get started.

Set your API key as an environment variable or create a .env file in your project directory:

ANTHROPIC_API_KEY=sk-ant-your-key-here

Using another provider? See the Models and Providers section to configure OpenAI, Ollama, Groq, etc.

3. Run

aru

That's it — aru is available globally after install.

Usage

Commands

Command	Description
Natural language	Just type — aru handles the rest
`/plan <task>`	Creates a detailed implementation plan
`/model [provider/model]`	Switch models and providers
`/mcp`	List available MCP servers and tools
`/commands`	List custom commands
`/skills`	List available skills
`/sessions`	List recent sessions
`/help`	Show all commands
`! <command>`	Execute shell commands
`/quit` or `/exit`	Exit aru

CLI Options

aru                                    # Start new session
aru --resume <id>                      # Resume session
aru --resume last                      # Resume last session
aru --list                             # List sessions
aru --dangerously-skip-permissions     # Skip permission prompts

Examples

aru> /plan create a REST API with FastAPI to manage users

aru> refactor the authentication module to use JWT tokens

aru> ! pytest tests/ -v

aru> /model ollama/codellama

Configuration

Models and Providers

By default, aru uses Claude Sonnet 4.6 (Anthropic). You can switch to any supported provider during a session with /model:

Provider	Command	API Key (`.env`)	Extra Installation
Anthropic	`/model anthropic/claude-sonnet-4-6`	`ANTHROPIC_API_KEY`	— (included)
Ollama	`/model ollama/llama3.1`	— (local)	`pip install "aru-code[ollama]"`
OpenAI	`/model openai/gpt-4o`	`OPENAI_API_KEY`	`pip install "aru-code[openai]"`
Groq	`/model groq/llama-3.3-70b-versatile`	`GROQ_API_KEY`	`pip install "aru-code[groq]"`
OpenRouter	`/model openrouter/deepseek/deepseek-chat-v3-0324`	`OPENROUTER_API_KEY`	`pip install "aru-code[openai]"`
MiniMax	`/model openrouter/minimax/minimax-m2.7`	`OPENROUTER_API_KEY`	`pip install "aru-code[openai]"`

To install all providers at once:

pip install "aru-code[all-providers]"

Ollama (local models)

To run models locally without an API key, install Ollama, start the server, and use any installed model:

ollama serve                    # Start the Ollama server
ollama pull codellama           # Download a model
aru                             # Start aru
# Inside aru:
/model ollama/codellama

Configuring the default model

You can set the default provider/model in aru.json so you don't need to switch manually every session:

{
  "default_model": "openrouter/minimax/minimax-m2.7",
  "model_aliases": {
    "minimax": "openrouter/minimax/minimax-m2.5",
    "minimax-m2.7": "openrouter/minimax/minimax-m2.7",
    "deepseek-v3": "openrouter/deepseek/deepseek-chat-v3-0324",
    "sonnet-4-6": "anthropic/claude-sonnet-4-6",
    "opus-4-6": "anthropic/claude-opus-4-6"
  }
}

The default_model field sets the main model. The model_aliases are shortcuts that can be used with /model <alias>.

Custom providers

You can configure custom providers with specific token limits:

{
  "providers": {
    "deepseek": {
      "models": {
        "deepseek-chat-v3-0324": { "max_tokens": 16384 }
      }
    },
    "openrouter": {
      "models": {
        "minimax/minimax-m2.5": { "max_tokens": 65536 },
        "minimax/minimax-m2.7": { "max_tokens": 131072 }
      }
    }
  }
}

Permissions (`aru.json`)

The aru.json file in the project root controls which shell commands aru can execute without asking for confirmation:

{
  "permission": {
    "allow": [
      "git *",
      "npm *",
      "pytest *",
      "python *",
      "uv run pytest *"
    ]
  }
}

Each entry is a glob pattern. Any command that doesn't match a listed pattern will prompt for confirmation before executing.

aru.json can also be placed at .aru/config.json.

AGENTS.md

Place an AGENTS.md file in your project root with custom instructions that will be appended to all agent system prompts.

`.agents/` Directory

.agents/
├── commands/       # Custom slash commands (filename = command name)
│   └── deploy.md   # Usage: /deploy <args>
└── skills/         # Custom skills/personas
    └── review.md   # Loaded as additional agent instructions

Command files support frontmatter with description and the $INPUT template variable for arguments.

MCP Support (Model Context Protocol)

Aru can load tools from MCP servers. Configure in .aru/mcp_config.json:

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-filesystem", "/allowed/path"]
    }
  }
}

Agents

Agent	Role	Tools
Planner	Analyzes codebase, creates structured implementation plans	Read-only tools, search, web
Executor	Implements code changes based on plans or instructions	All tools including delegation
General	Handles conversation and simple operations	All tools including delegation

Tools

File Operations

read_file — Reads files with line range support and binary detection
read_file_smart — Smart file reading focused on relevant snippets for the query
write_file / write_files — Writes single or batch files
edit_file / edit_files — Find-replace edits across multiple files

Search & Discovery

glob_search — Find files by pattern (respects .gitignore)
grep_search — Content search with regex and file filtering
list_directory — Directory listing with gitignore filtering
rank_files — Multi-factor file relevance ranking (name, structure, recency)

Code Analysis

code_structure — Extracts classes, functions, imports via tree-sitter AST
find_dependencies — Analyzes import relationships between files

Shell & Web

bash — Executes shell commands with permission gates
web_search — Web search via DuckDuckGo
web_fetch — Fetches URLs and converts HTML to readable text

Advanced

delegate_task — Spawns autonomous sub-agents for parallel task execution

Architecture

aru-code/
├── aru/
│   ├── cli.py              # Interactive CLI with streaming display
│   ├── config.py           # Configuration loader (AGENTS.md, .agents/)
│   ├── providers.py        # Multi-provider LLM abstraction
│   ├── agents/
│   │   ├── planner.py      # Planning agent
│   │   └── executor.py     # Execution agent
│   └── tools/
│       ├── codebase.py     # 16 core tools
│       ├── ast_tools.py    # Tree-sitter code analysis
│       ├── ranker.py       # File relevance ranking
│       ├── mcp_client.py   # MCP client
│       └── gitignore.py    # Gitignore-aware filtering
├── aru.json                # Permissions and model configuration
├── .env                    # API keys (not committed)
├── .aru/                   # Local data (sessions)
└── pyproject.toml

Built With

Agno — Agent framework with tool orchestration
Anthropic Claude — Sonnet 4.6, Opus 4.6, Haiku 4.5
tree-sitter — AST-based code analysis
Rich — Terminal UI
prompt-toolkit — Advanced input handling

Development

# Clone and install in editable mode with dev dependencies
git clone https://github.com/estevaofon/aru.git
cd aru
pip install -e ".[dev]"

# Run tests
pytest

# Run tests with coverage
pytest --cov=aru --cov-report=term-missing

Built with Claude and Agno

Metadata

Release files for aru-code 0.4.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for aru-code 0.4.0
File	Size	Uploaded
aru_code-0.4.0.tar.gz	120.5 kB	Mar 30, 2026	Details

Built distribution (wheel)

Table of built distributions (wheels) for aru-code 0.4.0
File	Interpreter	ABI	Platform	Reset
aru_code-0.4.0-py3-none-any.whl 72.9 kB Mar 30, 2026	Python 3	none	any	Details

Total release size: 193.4 kB

Release files / aru_code-0.4.0.tar.gz

Download URL	aru_code-0.4.0.tar.gz
Size	120.5 kB
Tags	Source
SHA-256 checksum How to use checksums	`7e72708415033a95881c2434319d07d63434d0422e8c2db2dd38ac58a3b31bdc`
BLAKE2b-256 checksum How to use checksums	`cafde5eb4b7d4b9d9112e8d6485c97ca88c9e020cc9ce40e06b375609d0abd5d`
Upload date	Mar 30, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`uv/0.7.8`

Release files / aru_code-0.4.0-py3-none-any.whl

Download URL	aru_code-0.4.0-py3-none-any.whl
Size	72.9 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`365df0492e89d14bc1195f76f08f3e06ea298ed6d1ff1fdfcda35de55aef14ef`
BLAKE2b-256 checksum How to use checksums	`f5f647c573c95663f62d0850d5f4daa4f825caedf5539f8aa71e39811856c69f`
Upload date	Mar 30, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`uv/0.7.8`