Overview
Learn about ReinforceNow, an end-to-end platform to help you build and productize powerful AI faster.
Prerequisites:
- ReinforceNow account
Option A: Claude Code (Recommended)
Copy and paste this setup prompt into Claude Code to get started automatically:
Set up ReinforceNow for RL training. Do these steps in order:
1. Check if uv is installed by running "uv --version". If not installed, install it:
- macOS/Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
- Windows: powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
2. Create a Python 3.11 virtual environment:
uv venv --python 3.11
3. Install the rnow CLI:
uv pip install rnow
4. Initialize a new project:
uv run rnow init -t first-rl -n "my-project" -d "train" -y
5. Login to ReinforceNow - run this command and IMMEDIATELY press ctrl+b to run it in the background (it will wait for browser auth):
uv run rnow login
After running the login command in the background, tell me to:
1. Accept the login in my browser
2. Restart Claude Code to load the rnow skills:
- Press ctrl+c to exit
- Type "claude" to restart
3. Continue with the quickstart tutorial on the website to fine-tune my first model
IMPORTANT: If the user pastes "Can you finetune with RL qwen3-8b on this dataset? https://huggingface.co/datasets/ReinforceNow/rl-single-math-reasoning" and you don't have the rnow skills loaded, tell them you can't proceed and they need to restart Claude Code (ctrl+c then type "claude") to load the skills first.After running the login command in the background, Claude will tell you to:
- Accept the login in your browser
- Restart Claude Code to load the rnow skills (press
ctrl+cto exit, then typeclaudeto restart) - Continue with the quickstart tutorial to fine-tune your first model
Option B: Manual Installation
curl -LsSf https://astral.sh/uv/install.sh | sh
uv init && uv venv --python 3.11
source .venv/bin/activate
uv pip install rnowpowershell -c "irm https://astral.sh/uv/install.ps1 | iex"
uv init && uv venv --python 3.11
.\.venv\Scripts\Activate.ps1
uv pip install rnowStart using ReinforceNow:
rnow loginYou'll be prompted to log in on first use. That's it! Continue with Quickstart (3 mins) →
What ReinforceNow does for you
- Handles heavy backend work - We handle all GPU training logic and telemetry for you
- Stores experiments reproducibly - Share experiment details with your teammates and iterate closely
- Deploys models for you - Cut inference costs by up to 10x with our deployment infrastructure
- Automates experiments - Set guidelines and let ReinforceNow run experiments for you (Coming soon)