Tailored AI

Getting Started

Prerequisites

  • Node.js 18+
  • One of the following LLM backends:
    • Ollama running locally
    • An OpenAI API key (or any OpenAI-compatible API)
    • An Anthropic API key

Hardware (for local LLMs)

Tested with an RTX 5090 (32GB VRAM) running models like devstral-small-2 (~15GB) and qwen3-coder:30b (~18GB).

Installation

Install globally via npm:

bash
npm install -g tailored-ai

Or with your preferred package manager:

bash
pnpm add -g tailored-ai
yarn global add tailored-ai

You can also install it as a project dependency:

bash
npm install tailored-ai

Setup

Create a config file and environment variables in your working directory:

bash
tai init                          # creates config.yaml and .env

Or manually:

bash
cp config.example.yaml config.yaml   # edit with your settings
cp .env.example .env                  # add your API keys

Quick start

Interactive REPL

bash
tai

Single message (non-interactive)

bash
tai -m "What is the current date?"

Use a named profile

bash
tai -p researcher -m "Find AI news"

JSON output for scripting

bash
tai -m "List files in /tmp" --json

Run as a service

Start the Discord bot, cron scheduler, and HTTP API together:

bash
tai --serve

Resume a previous session

bash
tai -s <session-id>

CLI options

| Flag | Short | Description | |------|-------|-------------| | --config <path> | -c | Path to config.yaml (default: ./config.yaml) | | --message <text> | -m | Send a single message and exit | | --session <id> | -s | Resume an existing session | | --profile <name> | -p | Use a named agent profile | | --json | -j | Output as JSON (for scripting) | | --serve | | Run as a service (Discord bot, cron, HTTP API) | | --help | -h | Show help |

Development

If you want to contribute or run from source:

bash
git clone https://github.com/quintushr/autonomous-agent.git
cd autonomous-agent
pnpm install
pnpm run build
pnpm run dev -- -m "Hello"