janito 2.3.0

A new Python package called janito.

Janito

Janito is a command-line interface (CLI) tool for managing and interacting with Large Language Model (LLM) providers. It enables you to configure API keys, select providers and models, and submit prompts to various LLMs from your terminal. Janito is designed for extensibility, supporting multiple providers and a wide range of tools for automation and productivity.

Features

• 🔑 Manage API keys and provider configurations
• 🤖 Interact with multiple LLM providers (OpenAI, Google Gemini, Mistral, DeepSeek, and more)
• 🛠️ List and use a variety of registered tools
• 📝 Submit prompts and receive responses directly from the CLI
• 📋 List available models for each provider
• 🧩 Extensible architecture for adding new providers and tools
• 🎛️ Rich terminal output and event logging

Advanced and Architectural Features

• ⚡ Event-driven architecture: Modular, decoupled system using a custom EventBus for extensibility and integration.
• 🧑‍💻 Tool registry & dynamic tool execution: Register new tools easily, execute them by name or call from automation pipelines.
• 🤖 LLM Agent automation: Supports agent-like workflows with the ability to chain tools or make decisions during LLM conversations.
• 🏗️ Extensible provider management: Add, configure, or switch between LLM providers and their models on the fly.
• 🧰 Rich tool ecosystem: Includes file operations, local/remote script and command execution, text processing, and internet access (fetching URLs), all reusable by LLM or user.
• 📝 Comprehensive event & history reporting: Detailed logs of prompts, events, tool usage, and responses for traceability and audit.
• 🖥️ Enhanced terminal UI: Colorful, informative real-time outputs and logs to improve productivity and insight during LLM usage.

Installation

Janito is a Python package. Since this is a development version, you can install it directly from GitHub:

pip install git+https://github.com/janito-dev/janito.git

First launch and quick setup

Janito integrates with external LLM providers (list below), and most of them require a subscription to get an API_KEY.

[!NOTE] Today, on June the 26th 2025, Google has a free tier subscription for its Gemini-2.5-flash model. Despite the limitation of the model and of the rate limit of the free tier, it can be used for testing janito. The API_KEY for Gemini is available here.

[!NOTE] Here a list of various services that provide free access or credits towards API-based LLM usage. Note that not all of them are supported by Janito, yet.

For a quick usage you can:

1. once you get the API_KEY from your favourite LLM provider, setup the API_KEY in Janito

janito --set-api-key API_KEY -p PROVIDER

2. then run janito from command line with the specific LLM provider of your choice

janito -p PROVIDER "Hello, who are you? How can you help me in my tasks?"

3. or you can run janito in interactive mode without the trailing argument

janito -p PROVIDER

4. if you want to setup a specific provider for any further interactions you can use:

janito -set provider=PROVIDER

[!WARNING] Currently the supported providers are: openai, google, azure_openai. You can get more details with janito --list-providers.

5. for more advanced setup, continue reading.

Usage

After installation, use the janito command in your terminal.

Janito has configuration options, like --set api-key API_KEY and --set provider=PROVIDER, that create durable configurations and single shoot options, like -p PROVIDER and -m MODEL, that are active for the single run of the command or session.

Basic Commands

• Set API Key for a Provider (requires -p PROVIDER)

  janito --set-api-key API_KEY -p PROVIDER
  

  Note: The -p PROVIDER argument is required when setting an API key. For example:

  janito --set-api-key sk-xxxxxxx -p openai
  

• Set the Provider (durable)

  janito --set provider=provider_name
  

• List Supported Providers

  janito --list-providers
  

• List Registered Tools

  janito --list-tools
  

• List Models for a Provider

  janito -p PROVIDER --list-models
  

• Submit a Prompt

  janito "What is the capital of France?"
  

• Start Interactive Chat Shell

  janito
  

Advanced Options

• Enable Inline Web File Viewer for Clickable Links

  By default, Janito can open referenced files in a browser-based viewer when you click on file links in supported terminals. To enable this feature for your session, use the -w or --web flag:

  janito -w
  

  This starts the lightweight web file viewer (termweb) in the background, allowing you to inspect files referenced in responses directly in your browser. Combine with interactive mode or prompts as needed.

  Tip: Use with the interactive shell for the best experience with clickable file links.

• Enable Execution Tools (Code/Shell Execution)

  By default, tools that can execute code or shell commands are disabled for safety. To enable these tools (such as code execution, shell commands, etc.), use the --exec or -x flag:

  janito -x "Run this code: print('Hello, world!')"
  

  Warning: Enabling execution tools allows running arbitrary code or shell commands. Only use --exec if you trust your prompt and environment.

• Set a System Prompt

  janito -s path/to/system_prompt.txt "Your prompt here"
  

• Select Model and Provider Temporarily

  janito -p openai -m gpt-3.5-turbo "Your prompt here"
  janito -p google -m gemini-2.5-flash "Your prompt here"
  

• Set Provider-Specific Config (for the selected provider)

  # syntax: janito --set PROVIDER.KEY=VALUE
  # example: set the default model for openai provider
  janito --set openai.model=gpt-4o
  

  Note: Use --set PROVIDER.key=value for provider-specific settings (e.g., openai.max_tokens, openai.base_url).

• Enable Event Logging

  janito -e "Your prompt here"
  

🌟 CLI Options Reference

Core CLI Options

Option	Description
-w, --web	Enable the builtin lightweight web file viewer for clickable file links (termweb).
--version	Show program version
--list-tools	List all registered tools
--list-providers	List all supported LLM providers
-l, --list-models	List models for current/selected provider
--set-api-key	Set API key for a provider. Requires -p PROVIDER to specify the provider.
--set provider=name	Set the current LLM provider (e.g., janito --set provider=openai)
--set PROVIDER.model=MODEL or --set model=MODEL	Set the default model for the current/selected provider, or globally. (e.g., janito --set openai.model=gpt-3.5-turbo)
-s, --system	Set a system prompt (e.g., janito -s path/to/system_prompt.txt "Your prompt here")
-r, --role	Set the role for the agent (overrides config) (e.g., janito -r "assistant" "Your prompt here")
-p, --provider	Select LLM provider (overrides config) (e.g., janito -p openai "Your prompt here")
-m, --model	Select model for the provider (e.g., janito -m gpt-3.5-turbo "Your prompt here")
-v, --verbose	Print extra information before answering
-R, --raw	Print raw JSON response from API
-e, --event-log	Log events to console as they occur
"user_prompt"	Prompt to submit for the non interactive mode (e.g. janito "What is the capital of France?")

🧩 Extended Chat Mode Commands

Once inside the interactive chat mode, you can use these slash commands:

📲 Basic Interaction

Command	Description
/exit or exit	Exit chat mode
/help	Show available commands
/multi	Activate multiline input mode
/clear	Clear the terminal screen
/history	Show input history
/view	Print current conversation history
/track	Show tool usage history

💬 Conversation Management

Command	Description
/restart	Start a new conversation (reset context)
/prompt	Show the current system prompt
/role <description>	Change the system role
/lang [code]	Change interface language (e.g., /lang en)

🛠️ Tool & Provider Interaction

Command	Description
/tools	List available tools
/termweb-status	Show status of termweb server
/termweb-logs	Show last lines of termweb logs
/livelogs	Show live updates from server log file
/edit <filename>	Open file in browser-based editor

📊 Output Control

Command	Description
/verbose	Show current verbose mode status
`/verbose [on	off]`

Extending Janito

Janito is built to be extensible. You can add new LLM providers or tools by implementing new modules in the janito/providers or janito/tools directories, respectively. See the source code and developer documentation for more details.

Supported Providers

• OpenAI
• OpenAI over Azure
• Google Gemini
• DeepSeek

See docs/supported-providers-models.md for more details.

Contributing

Contributions are welcome! Please see the CONTRIBUTING.md (if available) or open an issue to get started.

License

This project is licensed under the terms of the MIT license.

For more information, see the documentation in the docs/ directory or run janito --help.

---

Support

📖 Detailed Documentation

Full and up-to-date documentation is available at: https://janito-dev.github.io/janito/

---

FAQ: Setting API Keys

• Multiple API_KEY setup
• Use a specific model
• Fetch the availale LLM providers
• Fetch the availale models

Multiple API_KEY setup

To set an API key for a provider, you must specify both the API key and the provider name:

janito --set-api-key YOUR_API_KEY -p PROVIDER_NAME

You can have an API_KEY for each LLM provider

janito --set-api-key API_KEY_1 -p PROVIDER_1
janito --set-api-key API_KEY_2 -p PROVIDER_2

Then you can easily use one provider or the other without changing the API_KEY

janito -p PROVIDER_1 "What provider do you use?"
janito -p PROVIDER_2 "What provider do you use?"

If you omit the -p PROVIDER_NAME argument, Janito will show an error and not set the key.

Use a specific model

To use a specific model, you can use the -m option in the follwing way:

janito -m gpt-4.1-nano -p openai "What model do you use?"

Or you can use the durable --set option:

janito --set provider=openai 
janito --set model=gpt-4.1-nano
janito "What model do you use?"

Fetch the availale LLM providers

You can list all the LLM providers available using:

janito --list-providers

Fetch the availale models

Each LLM provider has its own models, the best way to check what are the available models is usign the following commands:

janito -p openai --list-models
janito -p google --list-models
janito -p azure_openai --list-models
janito -p deepseek --list-models

Ask Me Anything

When the FAQ are not enough, you can contact the contributors of the project by direct questions

Ask a question Read questions

Guidelines

• :mag: Ensure your question hasn't already been answered.
• :memo: Use a succinct title and description.
• :bug: Bugs & feature requests should be opened on the relevant issue tracker.
• :signal_strength: Support questions are better asked on Stack Overflow.
• :blush: Be nice, civil and polite.
• :heart_eyes: If you include at least one emoji in your question, the feedback will probably come faster.
• Read more AMAs
• What's an AMA?