Sign up at tryhamster.com. Once you're in, the path below tends to get teams to real value the fastest.
Hamster Studio helps you structure ideas, products, and the way your team operates — through Goals, Initiatives, Briefs, Blueprints, Methods, and the Context Graph. Every feature you'll use maps to one of those, so the rest of the product makes a lot more sense once the words are familiar. The Concepts section walks through them in about five minutes — it's worth doing before you go much further.
Hamster is most useful when it can read the work your team has already accumulated, so the next thing worth doing is connecting the tools you already use.
Most teams find GitHub and Slack are the two highest-leverage Connections to start with — GitHub because it's the largest single source of code-level context, Slack because it's where most of your team's running conversation already lives. From there, add Linear or Jira, Figma, Notion, Google Drive, and the Meeting Agent as the need comes up.
Each Connection widens what the AI can read when it works on a Brief, generates a Plan, or proposes an initiative. Two Connections is plenty to begin with; the rest tend to grow in naturally as you use Hamster more.
A Blueprint is the English description of something stable in your business — a product, a system, a team. The easiest way to start is to chat with Hamster about what you'd like a Blueprint for:
"Build me a Blueprint for our checkout product. The repo is
acme/checkout. It's used by first-time buyers."
Hamster reads what's already in your Context Graph, drafts a Blueprint, and hands you back something concrete to react to and edit. You can repeat this for the systems and surfaces that matter most — three or four Blueprints covering your highest-traffic products is usually plenty to ground every Brief that follows.
The fastest way to feel what Hamster actually does is to run an idea you've been kicking around through the whole loop, end-to-end.
A first run usually looks like this:
Doing this once on your own first lets you feel the loop — chat, Brief, Plan, delivered — before you bring anyone else into it.

Once the loop feels familiar, run it again with other people in the room. Invite teammates — Reviewer seats are free and unlimited, which makes them the right default for stakeholders who only need to read Briefs and vote on alignment.
The flow looks the same as solo: share a Brief, gather feedback, collect alignment votes, and deliver once everyone has signed off. The Blueprints and Briefs you've already built carry full context, so teammates joining a discussion can pick it up without needing a half-hour catch-up first.
If you ship code, you don't have to leave your editor. Connecting the CLI or the MCP Server lets your IDE — Cursor, Claude Code, or anything else that speaks MCP — pull a Brief, its Plan, the relevant Blueprints, and the connected context straight into your session. The Brief becomes the prompt; your Context Graph becomes the grounding; you write the code.
Cloud Agents are the right call when you want hands-off delivery. The CLI or MCP fit better when you want full control with full context. Both paths read from the same Briefs, Plans, and Blueprints, so you can pick whichever suits the work — and switch on a per-brief basis.