Get Started
with yaad

A CLI memory engine. One command to store, one to recall — in plain English, any time later. Works for you in the terminal and for your AI agents as a skill.

01

Install Ollama

yaad runs AI locally via Ollama — nothing leaves your machine. Install it, then pull the two models:

ollama pull mxbai-embed-large  # for search
ollama pull llama3.2:3b        # for recall (swap for any chat model)
02

Install yaad

curl -fsSL https://yaad.knl.co.in/install.sh | bash

Or grab a pre-built binary from GitHub Releases.

Then initialise config:

yaad config init
03

Save your first memory

Put all context in the content — the AI searches by meaning, so more detail means better recall later.

shell
yaad add "staging db is postgres on port 5433, host db.internal"
saved  [fact]
yaad add "deploy: run migrations first, restart workers, then clear cache"
saved  [note]
04

Recall with natural language

Ask in plain English — phrased however feels natural, not how you saved it.

shell
yaad ask "how do I connect to staging?"
postgres on port 5433, host db.internal
yaad ask "what's the deploy order?"
Run migrations first, restart workers, then clear cache.
05

Set a reminder

Add --remind with any natural language time. Surface it by hooking into your shell prompt:

shell
yaad add "submit PR for review" --remind "tomorrow 9am"
saved  [reminder] · due tomorrow 09:00
# add to ~/.bashrc or ~/.zshrc
export PROMPT_COMMAND="yaad check; $PROMPT_COMMAND"
# zsh
precmd() { yaad check }

06

Add to your AI agent

Install the Agent Skill once — your agent can save and recall memory across every session, automatically.

npx skills add kunalsin9h/yaad -g
claude code
/yaad "what's the staging db?"
postgres on port 5433, host db.internal
/yaad add "review PR #42" --remind "tomorrow 9am"
saved  [reminder] · due tomorrow 09:00

Your terminal and agent share the same store. Save in one, recall in the other.

What's next