rapt0r-cli 1.0.0

Offline terminal prompt builder — answer a few questions, get a perfect AI prompt in your clipboard. Zero tokens spent on explanations.

build fast. ship faster.
prompt builder by i73

---

Don't spend tokens explaining yourself — spend them on results.

RAPT0R is a terminal prompt builder. Answer a few questions, get a battle-tested AI prompt in your clipboard, or execute it directly against an LLM. Offline by design — your context window goes entirely to the actual work.

✨ Features

• 🔌 Offline by default — build complex prompts with zero API calls, or optionally stream execution directly to OpenAI/Anthropic.
• ⚡️ Fast — answer 3-5 questions, the finished prompt is already in your clipboard.
• 📁 Project context — drop a .rapt0r file in your repo and your stack/rules/architecture get injected into every prompt automatically.
• 🔎 Instant search — press f anywhere to fuzzy-search templates across all categories.
• 🌳 Branching — smart templates ask follow-up questions based on your answers.
• 🛠 Built-in manager — create, edit, and delete templates right from the menu.
• 🎨 Polished TUI — custom colors, number hotkeys, icons, built on InquirerPy and rich.

---

🚀 Install

pip install rapt0r-cli

Run it:

rapt0r

Or run from source

git clone https://github.com/i73i73i73i73/RAPT0R.git
cd RAPT0R
pip3 install -e .
rapt0r          # or: python3 main.py

---

📁 Project context (.rapt0r)

Stop re-explaining your project to the AI in every prompt. Describe it once — RAPT0R appends it to every prompt it builds.

Create a .rapt0r file in your project root (any plain text or markdown):

Stack: FastAPI + React + PostgreSQL
Style: type hints everywhere, no classes where functions do
Architecture: monorepo, backend in /api, frontend in /web
Rules: never touch migrations by hand, tests are pytest

Run rapt0r from anywhere inside the project — it finds the file by walking up the directory tree and adds a ## Project context section to the generated prompt.

---

⌨️ Hotkeys

Key	Action
1-9	Jump to item
f	Fuzzy search across all templates
0	Exit / back
Ctrl+C	Cancel current flow

---

📝 Writing your own templates

Templates are plain Markdown with YAML frontmatter. Press Create template in the menu, or drop a .md file into a category folder:

---
name: "✨ My awesome prompt"
questions:
  role: "What role should the AI play?"
  task: "Describe the task in one sentence:"
branch:
  question: "Add a constraints section?"
  y:
    questions:
      limits: "List the constraints, comma-separated:"
    body: "constraints_body"
---

You act as: {{role}}.
Your task: {{task}}.

{{constraints_body}}

Syntax

• name — display name in the menu.
• questions — key: "Question?" pairs; keys become {{variables}} in the body. Add choices: for a pick-list.
• branch (optional) — yes/no fork that asks extra questions and switches the body.
• {{variable}} — replaced with your answers.

Where templates live: the repo's rapt0r/templates/ when running from source, ~/.rapt0r/templates/ when installed via pip (seeded automatically on first run). Generated prompts are saved to ~/.rapt0r/prompts/ (or prompts/temp/ in a source checkout).

---

🧠 Philosophy

Precise context = fewer tokens = better results.

Every API call where you explain your project from scratch costs money and context window. RAPT0R forces the discipline of giving the AI exactly the inputs it needs — offline, for free, in seconds.

_{built with ❤️ for AI engineering}