HypAware records sessions from the AI agents your team runs and connects to the
tools around them. hyp setup detects the agents installed on a host and offers
to record them; hyp attach <client> adds one later. Each integration's page
covers what is recorded, how to set it up, and its limits.
Setup also installs HypAware's skills in each recorded agent that has a skills
directory, so the agent can query its own history. See
clients and history for attaching, history imports, and detaching
in general.
Status
Status
Meaning
Supported
Built and maintained by the HypAware team. Setup follows each integration's guide.
Beta
Built by the HypAware team and usable today, but it depends on an interface that may change. Behavior can change between releases.
Community
Built and maintained outside the HypAware team, usually as a plugin. No community integrations are listed yet.
Runs on is where the agent or tool itself runs. HypAware runs on a host you
control: install it next to an agent on a server or CI runner, and for cloud
agents the collection route depends on the provider. Codex Cloud uses your
existing local login; the tested Claude Code Cloud recipe runs the
collector inside the cloud environment.
Native hooks signal local transcript changes, plus recovery every five minutes; no billed token usage or cost
Yes
droid
Cloud agents
For Anthropic-managed sessions, see the
Claude Code Cloud recipe, tested with
HypAware 1.42.0. It includes setup, session startup, and upload after replies.
The collector runs without a browser or service manager on factory hosts, containers, and CI runners such as GitHub Actions, and records the agents the job runs