← Back to Projects

Memora

Memory-augmented CLI that learns from your workflow and retrieves past shell, Python, and SQL actions with semantic search.

June–September 2025 Solo Developer active
python cli llm nlp
GitHub →

Technology

  • Python
  • OpenAI Embeddings
  • Click
  • SQLite

What it is

A lightweight CLI REPL that records commands across shell, Python, and SQL, embeds them, and suggests relevant past actions on demand.

Why it matters

  • Reduces command lookup and context switching.
  • Captures reusable workflows that usually disappear in shell history.
  • Works fully offline with local storage.

How it works

  • Semantic retrieval: embeddings for intent-level similarity, not string match.
  • Context adapters: shell, !py for Python, !sql for SQL.
  • Personalization: tracks frequency and outcomes to rank suggestions.
  • Local DB: SQLite for fast reads, no cloud dependency.

Example

? how do I list all files including hidden
→ Suggestion: ls -a

Tech

  • Language: Python
  • CLI: Click
  • Vector search: OpenAI embeddings
  • Storage: SQLite

My role & links

  • Designed the data model, retrieval logic, and REPL UX.
  • Built adapters for shell, Python, and SQL.
  • Code: GitHub