Norma, in your
Claude workspace.
The same agents that run by email, reachable inside Claude as a connector — same firm library, same memory, same tools, in the chat your studio already uses.
Same library. Two doors.
Email is the primary surface — where most studios live. Claude is the same Norma, reached through a connector. Whichever door you open, you're working the same firm library.
Forward and reply.
CC an agent on a thread, forward a quote, reply to a spec request. Norma works the way your team already runs projects — no new app to open, no behavior to change.
Ask in chat.
Add Norma as an MCP connector and the same agents answer inside Claude. Native to studios already drafting and researching there — no inbox round-trip.
What designers do with it.
Plain requests, in the chat window. Norma sources the answer against your firm's library and applies the result with citations and receipts — everything reversible, and the judgment calls staged for you.
- 01
Parse this Steelcase quote into the Tampa refresh project.
→ Line items extracted, matched to the library, and filed on the project — cited, receipted, revertible.
- 02
What lounge chair did we spec on the last hospitality lobby?
→ Pulled from the firm library, with the project, date, and source cited.
- 03
Find three substitutes for the Andreu World Solo under $1,800/seat.
→ Sourced from the library, priced and compared, staged for your call.
Email is still the front door.
You don't need this page to use Norma. Most teams never leave their inbox.
Most studios only ever need email. It's where the work already happens — quotes arrive at edgewood@agent.norma.llc, threads run, approvals get sent. Norma fits inside that, and that's the whole product for the majority of teams.
Claude is for Claude-native studios — teams already drafting, researching, and thinking in the chat. For them, an MCP connector removes the inbox round-trip.
Either way it's the same agents and the same firm library underneath. Email or Claude is a preference, not a different product.
Wiring Norma into your own agents instead of Claude's? That is a different product — Norma API, where you bring your own harness and tokens.
Check whose Claude you're on. Norma's side of every turn runs as API traffic under Norma's account — Anthropic never trains on it (see privacy). The chat around the connector, though, runs on your Claude plan. On Claude for Work — Team or Enterprise — Anthropic doesn't train on your content by default. On personal plans (Free, Pro, Max), your chats may be used for model training depending on the choice in your privacy settings. For firm work: use a Work plan, or switch that setting off.
Open the door your studio prefers.
Start with Google — your agent is minted in a minute, and you connect Claude from your profile once it exists. Same Norma underneath.