Skip to content

Quick Start

This guide gets you up and running with janito in minutes.

1. Configure Your Settings

The first time you run janito, you'll need to configure your API settings. You can do this interactively or with command-line flags.

Interactive Configuration

Run the configuration wizard:

janito --config

You'll be prompted for:

Setting Description Example
Provider Your API provider, selected from a list openai, custom
API Key Your API key (masked for security) sk-xxxxxxxxxxxxxxxx
Model Model name to use gpt-5.6-luna, claude-sonnet-5
Max Output Tokens Maximum output tokens (default: 65536) 65536
Max Input Tokens Maximum input tokens / context window (default: 128000) 128000

Quick Configuration with Flags

Set options directly from the command line. Note that --set and --set-api-key must be used in separate commands:

# OpenAI example (two steps: config, then API key)
janito --set provider=openai --set model=gpt-5.6-luna
janito --set-api-key="sk-your-key" --provider openai

# Local LLM example (local servers use the `custom` provider + endpoint)
janito --set provider=custom --set endpoint="http://localhost:1234/v1" --set model="local-model"
janito --set-api-key="not-needed" --provider custom

2. Run Your First Prompt

Single Prompt

janito "What is the capital of France?"

Interactive Chat

janito

Type your messages and press Enter. Commands:

Command Description
/exit End the session
clear Clear conversation and start a new one
Ctrl+D / Ctrl+Z Exit the shell (EOF)
/help List all commands (with descriptions), tool modes and shortcuts

For the full list of chat commands, see Interactive Mode.

Pipe Input

echo "Tell me a joke" | janito

3. View Your Configuration

janito --show-config

Examples

OpenAI

# Step 1: Set provider and model
janito --set provider=openai --set model=gpt-5.6-luna
# Step 2: Store API key
janito --set-api-key="sk-your-key" --provider openai
# Step 3: Run prompt
janito "Explain quantum computing"

Local LLM (LM Studio, Ollama)

Local servers are reached through the custom provider, which requires an endpoint (--set endpoint=...):

# Step 1: Set provider, endpoint, and model
janito --set provider=custom \
        --set endpoint="http://localhost:1234/v1" \
        --set model="local-model"
# Step 2: Store placeholder API key
janito --set-api-key="not-needed" --provider custom
# Step 3: Run prompt
janito "What is 2+2?"

Custom Provider

# Step 1: Set provider, endpoint, and model
janito --set provider=custom --set endpoint="http://localhost:8000/v1" --set model="my-model"
# Step 2: Run prompt
janito "Hello"

What's Next?