my-ai-cli

A production-grade CLI for interacting with OpenAI's API, with optional RAG (Retrieval-Augmented Generation) support for document-grounded answers.

Features

• Single-question mode — myai ask "What is Python?"
• Interactive chat — myai chat
• File-aware answers — myai ask --file doc.txt "Summarise this"
• RAG support — ingest documents and query them semantically
• Streaming responses — tokens appear as they are generated
• Retry with backoff — handles rate limits gracefully
• Model selection — --model gpt-4o

Quick Start

# 1. Clone and install
git clone <repo-url>
cd my-ai-cli
python -m venv venv && source venv/bin/activate
pip install -e .

# 2. Set your API key
cp .env.example .env
# Edit .env and add your OPENAI_API_KEY

# 3. Ask a question
myai ask "Explain async/await in Python"

# 4. Start a chat session
myai chat

# 5. Ingest documents and query with RAG
myai ingest report.pdf notes.txt
myai ask --rag "What did the report say about Q3?"

Commands

myai ask

myai ask [OPTIONS] QUESTION

Options:
  -f, --file PATH      Append file contents to the question
  -m, --model TEXT     OpenAI model (default: gpt-4o-mini)
  --no-stream          Disable streaming output
  --rag                Use local RAG vector store
  --rag-dir PATH       Vector store directory (default: ~/.myai/store)
  -v, --verbose        Show retrieved RAG context

myai chat

myai chat [OPTIONS]

Options:
  -m, --model TEXT     OpenAI model
  --rag                Use RAG context for each turn
  --rag-dir PATH       Vector store directory
  -v, --verbose        Show retrieved context

Special commands (inside chat):
  /quit | /exit        End the session
  /clear               Reset conversation history
  /help                Show available commands

myai ingest

myai ingest [OPTIONS] PATHS...

Options:
  --store-dir PATH     Where to save the FAISS index (default: ~/.myai/store)
  --chunk-size INT     Characters per chunk (default: 1000)
  --overlap INT        Overlap between chunks (default: 200)

Supported formats: .txt, .md, .markdown, .pdf

Configuration

Priority: environment variables > ~/.myai/config.yaml > defaults.

Variable	Default	Description
OPENAI_API_KEY	—	Required
OPENAI_MODEL	gpt-4o-mini	Chat model
OPENAI_EMBEDDING_MODEL	text-embedding-3-small	Embedding model
OPENAI_MAX_TOKENS	2048	Max response tokens
OPENAI_TEMPERATURE	0.7	Sampling temperature
LOG_LEVEL	INFO	Logging verbosity

Example ~/.myai/config.yaml:

model: gpt-4o
temperature: 0.3
max_tokens: 4096

Development

pip install -r requirements-dev.txt
pytest --cov=src tests/
black src/ tests/
ruff check src/ tests/