Your browser should remember your work

Shaw keeps context from your tasks and brings it back when your agent needs it.

Your browsing history becomes memory

Shaw knows which sites to use for which work. It turns your browsing history into memory, so you don't have to repeat context every time.

Find the candidate I opened yesterday and prep interview notes

After each task, Shaw saves what matters. It gets better at your work over time.

Dreaming Shaw extracts key details from each task, like people, companies, and decisions.

Site-specific memory When the agent gets stuck on a website, Shaw remembers the path that worked.

Unlimited context Shaw keeps context across tasks. Your agent understands past work and continues from there.

Memory is written in markdown and stored on your device. You can see where it came from. If something is wrong, edit it.

Shaw |Main Agent·History·Memory·Agents

Memory History

Last dream 19 hours ago · 1 session(s) since

Extracted from session

session_completed · 18.4s · 202743 tokens · cached 74%

1 file changed+6-0

episodic/2026-06-21.md+6-0

You are now acting as the memory extraction subagent.

You're running asynchronously to extract episodic memory from this conversation.

Analyze the most recent ~20 messages above and use them to update your episodic memory.

Available tools: read_file, write_file, edit_file, memory_search, get_time, and read-only Bash.