rnow-cli

ReinforceNow CLI Reference

The rnow CLI manages RLHF training projects on the ReinforceNow platform.

Installation

pip install rnow

Command Overview

rnow login

Authenticate using OAuth device flow.

rnow login [OPTIONS]

Example:

rnow login
# Opens browser for authentication
# Stores credentials in ~/.reinforcenow/credentials.json

rnow logout

Remove stored credentials.

rnow logout

rnow status

Check authentication status and running jobs.

rnow status

Output:

Logged in as: user@example.com
Organization: My Team (org_abc123)
Active runs: 2
  - run_xyz789 (running) - Math Training
  - run_def456 (queued) - Code Agent

rnow orgs

List or select organizations.

# List all organizations
rnow orgs

# Select an organization
rnow orgs ORG_ID

Example:

rnow orgs
# Output:
# * org_abc123 - My Team (owner)
#   org_def456 - Other Team (member)

rnow orgs org_def456
# Switched to: Other Team

rnow init

Initialize a new project from a template.

rnow init [OPTIONS]

Available Templates

Examples:

# Create SFT project
rnow init --template sft --name "my-sft-project"

# Create RL project with tools
rnow init --template rl-tools

# Create from tutorial
rnow init --template tutorial-reward

Generated Files

rnow run

Submit project for training.

rnow run [OPTIONS]

Required files:

• config.yml - Configuration
• train.jsonl - Training data
• rewards.py - Reward functions (RL only)

Optional files:

• tools.py - Tool definitions
• requirements.txt - Python dependencies

Example:

cd my-project
rnow run

# Output:
# Validating project...
# Uploading files...
# Starting run: run_abc123xyz
# View at: https://www.reinforcenow.ai/runs/run_abc123xyz

rnow stop

Cancel an active training run.

rnow stop RUN_ID

Example:

rnow stop run_abc123xyz
# Are you sure you want to stop run_abc123xyz? [y/N]: y
# Run stopped.
# Duration: 2h 15m
# Cost: $12.50

rnow test

Test RL rollouts locally before submitting.

rnow test [OPTIONS]

Available Models

OpenAI API models (default, fast):

• gpt-5-nano - Fastest, recommended for quick testing
• gpt-5-mini - Faster
• gpt-5.2 - Balanced
• gpt-5-pro - Highest quality

GPU models (slower, uses actual training infrastructure):

Text models:

• Qwen/Qwen3-8B
• Qwen/Qwen3-32B
• Qwen/Qwen3-30B-A3B
• Qwen/Qwen3-235B-A22B-Instruct-2507
• meta-llama/Llama-3.1-8B-Instruct
• meta-llama/Llama-3.3-70B-Instruct
• deepseek-ai/DeepSeek-V3.1

VLM models (for vision tasks with screenshots):

• Qwen/Qwen3-VL-30B-A3B-Instruct - Vision-language model
• Qwen/Qwen3-VL-235B-A22B-Instruct - Larger VLM

Important: Default gpt-5-nano is text-only and cannot process images. For VLM projects that return screenshots (e.g., browser agents), use --model Qwen/Qwen3-VL-30B-A3B-Instruct to test with actual vision capabilities.

Examples

Basic test:

rnow test
# Runs 1 rollout with gpt-5-nano (text-only)

Multiple rollouts:

rnow test -n 5

Test specific entries:

rnow test --entry 0,3,7
# Tests entries at indices 0, 3, and 7 from train.jsonl

Test with VLM model (for vision/screenshot projects):

rnow test --model Qwen/Qwen3-VL-30B-A3B-Instruct -n 1
# Uses actual VLM that can see images

Test with GPU model:

rnow test --model Qwen/Qwen3-8B -n 3
# Uses GPU infrastructure instead of OpenAI API

Test Output

Rollout 1/3
Entry: 0
Prompt: What is 2+2?

Turn 1:
  Assistant: The answer is 4.

Rewards:
  accuracy: 1.0
  format_check: 1.0
Total: 1.0

---
Rollout 2/3
...

rnow download

Download a trained model checkpoint.

rnow download RUN_ID [OPTIONS]

Example:

rnow download run_abc123xyz -o ./my-model
# Downloading checkpoint...
# Progress: 100%
# Saved to: ./my-model/