Quick Start

Learn how Kyew works in 5 minutes: teach your AI what to remember, let it learn your patterns, and build tools you can use every day.

How Kyew Learns

Kyew follows a simple cycle that makes your AI smarter over time:

You work → Kyew remembers → Patterns emerge → Your AI improves

1. You work: Use your AI as you normally do
2. Kyew remembers: Important context is saved across sessions
3. Patterns emerge: Kyew notices what you repeat
4. Your AI improves: Saved knowledge is applied automatically in future conversations

Step 1: Teach Your AI

As you work, tell your AI to remember things that matter:

"remember that I prefer weekly reports sent on Monday mornings"

Your AI saves this with context so it can be recalled later:

"remember that our team standup is at 9am and we track tasks in Linear"

What to Remember

Good memories are specific and useful:

• "I use the company Slack channel #updates for announcements"
• "My preferred meeting length is 30 minutes, not 60"
• "Our quarterly reports pull data from the finance spreadsheet in Google Drive"

Avoid things that are too vague to act on:

• "The meeting went well"
• "I did some work"

Step 2: Your AI Remembers

Later, your AI can recall what you've taught it:

"what do you know about my meeting preferences?"

Your AI searches its memory and finds what's relevant:

"what do you remember about our team processes?"

Step 3: Patterns Emerge

After a few conversations, Kyew notices what you keep repeating:

"analyze patterns in my workflow"

Kyew finds recurring themes across your conversations:

"I've noticed you describe your reporting preferences in 3 of your
last 5 sessions. Want me to save this as a reusable workflow?"

Step 4: Save as a Skill

Turn a pattern into a reusable skill your AI can apply automatically:

"save that as a skill for my weekly reporting workflow"

Kyew creates a draft for you to review:

Skill created: "Weekly Reporting Workflow"
Status: Draft (awaiting your approval)

Step 5: Approve and Activate

Review pending skills and approve the ones you want:

"show pending skills"

Then approve:

"approve the weekly reporting skill"

Once approved, your AI will use this skill automatically when the context matches.

Step 6: Your AI Applies What It Knows

When you work on similar tasks, your AI already has the context it needs:

"help me prepare this week's report"

Your AI recalls your preferences and applies the relevant skill automatically — no re-explaining needed.

Complete Example

Here's what a typical week with Kyew looks like:

# Monday: Starting a new project
You: remember that our client reports go to sarah@company.com and need the Q4 metrics

# Wednesday: Working on a presentation
You: remember that we always use the blue brand template for client decks

# Friday: Preparing a summary
You: remember that weekly summaries should include hours billed and tasks completed from Linear

# Next Monday: Analyze patterns
You: analyze patterns in my client-reporting workflow

# Result: Kyew notices you have 3 related memories about client reporting

# Generate a skill
You: generate a skill from those client reporting memories

# Review and approve
You: show the pending skill
You: approve it

# Future use — your AI already knows the drill
You: help me prepare this week's client report
# → Your AI already knows the recipient, template, and data sources

Tips for Getting the Most Out of Kyew

Be Specific

The more detail you give, the more useful the memory:

# Good
remember that client reports go to Sarah

# Better
remember that weekly client reports go to sarah@company.com every Friday,
using the blue brand template, with hours from Linear and metrics from the dashboard

Group Related Things Together

Use consistent topics so Kyew can find patterns:

remember in "client-reporting" that reports go to sarah@company.com
remember in "client-reporting" that we use the blue brand template
remember in "client-reporting" that data comes from Linear and the dashboard

Review Before Approving

Always review generated skills before approval. Kyew is good at synthesizing patterns, but you know your workflow best.