Quickstart

1. Install

npm i -g hypaware
hyp setup

Install HypAware using npm i -g hypaware, then run hyp setup --guide and follow its instructions.

Setup walks you through it: pick the tools to record, install the daemon, and attach the tools you picked. When it finishes it shows you what it has already found on the machine.

2. Check it is capturing

hyp status

Run hyp status and tell me whether HypAware is installed and recording correctly, and what to fix if not.

This is the answer to "is HypAware working." It shows the daemon, what is being recorded, which tools are attached, the retention window, and anything that needs attention. Add --json for machine-readable output.

3. Look at what you already have

hyp query overview

Use HypAware to tell me what is recorded on this machine: which datasets, how fresh, and a summary of models, tokens, and tools.

Your history was already on the machine. This summarizes it: sessions, tools, models, and tokens over the last 30 days.

4. Ask it a question

hyp ask "From my HypAware history, which repo used the most tokens this month, and which model?"

From my HypAware history, which repo used the most tokens this month, and which model?

To find out what repos are consuming the most tokens this month.

hyp ask "From my HypAware history, where did my agents go off the rails recently?"

From my HypAware history, where did my agents go off the rails recently?

To see where agents got stuck.

hyp ask "From my HypAware history, which tools failed most often this week?"

From my HypAware history, which tools failed most often this week?

To find where the agents are having repeated tool failures.

5. Generate a report

A report is your agent reading your recent sessions and writing up what it found: usage, the work itself, where sessions broke down, and specific changes to make, with the sessions behind each one. It runs in Claude Code or Codex, so one of them needs to be recorded. Tell it the period and what to look into:

hyp report generate "Cover last week and focus on repeated debugging work"

Use the hypaware-report skill to cover last week and focus on repeated debugging work.

To see where your agents kept fixing the same problems.

hyp report generate "Cover the last two weeks and focus on tool failures and retries"

Use the hypaware-report skill to cover the last two weeks and focus on tool failures and retries.

To find the tools and commands that keep failing, and what to change.

hyp report generate "Give me one recommendation about repeated test failures"

Use the hypaware-report skill to give me one recommendation about repeated test failures.

To get a single recommendation instead of a full report.

With no instructions, a report covers the previous calendar month. Finished reports are saved on this machine, and hyp report list shows them. To share one with your team, connect first (step 7), then publish it with its saved name and the period it covers, such as 2026-W41:

hyp report publish REPORT_NAME --kind usage-review --period PERIOD --remote team

Use HypAware to publish the report folder REPORT_NAME to our organization as a usage review for PERIOD.

Published reports appear in your organization’s HypAware Cloud dashboard, where your team can read them. Read the report for anything private before you publish it.

6. Fix a recommendation

In the repository a recommendation applies to, run hyp report fix. It lists your published reports and their recommendations, and starts your agent on the one you pick. To fix a single recommendation, publish its page first with hyp report recommend.

hyp report fix

Use HypAware to list my open recommendations with hyp report list --status open and ask me which one to fix. Read the one I pick with hyp report get and apply it in this repository. When it is landed, mark it applied with the pull request link.

To go straight to one recommendation, pass its ID. hyp report list prints the ID under each report:

hyp report fix hyprec-0123456789abcdef

Read recommendation hyprec-0123456789abcdef with hyp report get and apply it in this repository. When it is landed, mark it applied with the pull request link.

When the change lands, your agent marks the recommendation applied, with the pull request link.

7. Connect to your team

hyp remote login

Works whether you have an account or not. It opens a browser sign-in, enrolls this machine with your organization, and starts forwarding what it records. From then on you can query across every connected machine, not only your own:

hyp ask "From our HypAware team history, which of our sessions this week reached a pull request?"

From our HypAware team history, which of our sessions this week reached a pull request?

To find out where your software factory is shipping.

One person can sign in and sync on their own. To add teammates to your organization, contact us at hypaware.ai/contact.

8. Control what is recorded

Mark any folder as ignore (never recorded), local-only (recorded, never synced), or sync, with a .hypignore file in the repo or with the hyp privacy commands. To be asked before each new folder is synced:

hyp privacy folders ask

To keep one conversation out of the record, hyp session ignore. To delete local recordings, hyp privacy purge. To stop syncing this machine, hyp leave.

9. Add or remove a tool

hyp attach claude

Attaching is idempotent; run it again and nothing changes.

hyp detach claude

Removes only the settings HypAware added. Then hyp status to confirm.

Where to go next

The full reference, every command and every option, is in the docs. If something in this guide didn't work the way it says, talk to an engineer; that is exactly the kind of thing we want to hear.