Skip to content

START_HERE.md

Latest commit

聽

History

History
130 lines (85 loc) 路 4.88 KB

START_HERE.md

File metadata and controls

130 lines (85 loc) 路 4.88 KB
Raw

Start Here

One sentence: R.A.I.N. Lab is an AI research assistant that helps you explore sound, resonance, and physics ideas without rediscovering things you already know.

Why you need it: You're doing research with AI. This tool makes sure your AI doesn't waste time finding "new" ideas that you already knew or that are widely known.

If You Only Read One Thing

Install once, then use this command every day:

python rain_lab.py

For non-technical users, this is the only daily entry point you need. The script opens a guided wizard, pressing Enter starts the no-setup instant demo, and every beginner/demo run updates a local showcase page in meeting_archives/RAIN_LAB_SHOWCASE.html. Those runs also generate a screenshot-friendly HTML share card plus a matching poster SVG in meeting_archives/.

Windows users should start with .\INSTALL_RAIN.cmd. macOS/Linux users should start with ./install.sh.

You can safely ignore files like rain_unique.py, james_reader.py, and other specialized scripts unless a maintainer tells you to use them. chat_with_james.py is used for the first installer handoff, but it is not the main product launcher.

Stable Path

Start with python rain_lab.py. That launcher, plus the default Rust build, is the stable core.

Integrations such as Matrix, Lark, Nostr, WhatsApp Web, hardware, and experimental workflow automation are opt-in extensions. Treat them as add-ons, not the baseline product.

If you are deciding whether a file, script, or integration belongs in the default path, use the product boundary: docs/project/product-boundary.md.

If you need the full support boundary, see docs/project/stability-tiers.md.

Name Guide (Plain English)

  • R.A.I.N. Lab: the product experience you use.
  • ZeroClaw: the Rust runtime engine under the hood.
  • James Library: the Python workflow collection in this repository.
  • Vers3Dynamics: the project/organization branding.

If you're just using the tool, think of all of this as one app and start with python rain_lab.py.

Checklist

Before first use, make sure you have one supported install route:

  1. Windows: run .\INSTALL_RAIN.cmd
  2. macOS/Linux: run ./install.sh
  3. Optional local models: install Ollama or LM Studio if you want local inference instead of the instant demo or a hosted provider
  4. Optional hosted models: paste an API key into the bootstrap prompt so .env is created for you

If you are unsure whether setup is complete, run:

python rain_lab.py --mode validate

First-Time Onboarding Flow

For a new non-technical user:

  1. Run .\INSTALL_RAIN.cmd on Windows, or ./install.sh on macOS/Linux.
  2. If the installer hands you off to James, use that as the welcome screen, then return to python rain_lab.py for the main product workflow.
  3. Run python rain_lab.py and press Enter for the instant demo, or choose Beginner mode.
  4. If you want to wire up local or hosted models, run python rain_lab.py --mode first-run.
  5. After a session, open meeting_archives/RAIN_LAB_SHOWCASE.html to revisit recent runs, poster previews, and copy the next commands.
  6. If anything fails, run python rain_lab.py --mode validate.

One Command to Start

Any System (Recommended)

python rain_lab.py

That's it. Just run that command, press Enter for the instant demo if you want the fastest path, and use the generated showcase page and poster-style share outputs to keep going.

What Can You Do?

When you want to... Run this
I'm not sure where to start python rain_lab.py (starts wizard)
Give it one idea and let it choose for me python rain_lab.py --mode beginner --topic "your idea"
Try a no-setup instant demo python rain_lab.py --mode demo --preset startup-debate
Chat with AI about my research python rain_lab.py --mode chat --topic "your topic"
Check if my system is ready python rain_lab.py --mode validate
See what AI models are available python rain_lab.py --mode models
Set everything up for the first time python rain_lab.py --mode first-run
Run a structured research meeting python rain_lab.py --mode rlm --topic "your topic"

Quick Troubleshooting

"Python not found"

"Ollama not found"

  • Install Ollama only if you want local-model inference. Otherwise press Enter for the instant demo or use a hosted API key in .env.

Not sure what to do?

  • Just run python rain_lab.py and it will ask you what you want to do

Need Help?

  • Simplest start: Run python rain_lab.py and choose from the menu
  • Install routes: See README.md
  • Technical details: See README.md
  • Problems?: Try python rain_lab.py --mode validate