Build with Emily.
What you get without building it.
Most of the work in a useful AI assistant is memory: what to keep, what to drop and how to stay consistent. Emily does that part.
Memory you do not have to build
Emily keeps context across conversations, so your app picks up where the last session left off instead of starting from zero.
Learning with guardrails
Before anything is stored, three protocols remove duplicates, block unreliable thoughts and protect her identity.
How the Epistemic Firewall worksMemory you can inspect
In Emily Brain, people can see what Emily remembers, how memories connect, and ask her why she holds one.
The model you choose
Emily works with leading AI models, so you are not tied to a single provider.
AI models Emily can run on
How Emily behaves in your app.
First conversations
Emily starts learning what matters to the people using your app. Only information that passes her three checks becomes a memory.
Over weeks and months
Conversations build on each other. Emily recalls earlier decisions and context without your app resending them.
At any time
People can open a memory and ask Emily why she holds it, what it connects to and how confident she is.
Web, email, MCP or API.
Start with the one that matches how your people work. They all reach the same Emily and the same memory.
Web
Chat with Emily in the browser, with voice and files. Register with Google; no code needed.
Create an accountMCP
Connect compatible AI tools to Emily, so they share her memory.
Ask for the MCP connection guideEmail Emily or forward a thread, and she replies with the context she already holds.
Ask for the email setup guideCompanion Developer: $179 CAD a month, coming soon
The planned developer tier bundles MCP, the API and your own key, with 25,000,000 tokens monthly. Planned pricing, not yet available to purchase.
EmilyRemote: the CLI and SDK.
EmilyRemote signs you in to your own Emily and streams conversations, from a terminal or from JavaScript and TypeScript. Sign-in is OAuth device authorization with browser approval, handled by Robson.
1. Install and sign in
npm install -g @robsonai/emilyremote
emilyremote setup
emilyremote self-check
emilyremote login --email you@example.com
emilyremote ask "What should I focus on today?"Open the approval link and approve with the same account. On a headless server, add --no-browser --credential-store file and approve on another computer.
2. Call Emily from your code
import { EmilyClient } from '@robsonai/emilyremote';
const account = new EmilyClient();
const signedIn = await account.login({ headless: true,
onEvent(e) {
if (e.type === 'login_required')
console.error(e.verification_url, e.user_code);
} });
const personal = account.withProfile(signedIn.profile);
for await (const e of personal.ask({ text: 'Help me plan today.' })) {
if (e.type === 'result') console.log(e.result.response_text);
}ES modules and CommonJS, with TypeScript declarations. Credentials stay in the native credential store; the API never returns a token.
Requirements: Node.js 22 or later. Runtimes included for Linux x64 (glibc 2.38+) and Apple Silicon (macOS 26+). Intel Mac, Linux ARM and Windows are not supported yet.
@robsonai/emilyremote on npmThe inbox you already use.
Email is how most work arrives, so it is a way into Emily that needs nothing new. Send her a question or forward a thread, and she answers with the context she already holds.
- What she learns from an email joins the same memory as your chats, MCP tools and apps
- Pick up on the web where an email thread left off, without re-explaining
- Works from any mail client, on any device
When paid plans launch, email access is planned as the Email to Clone add-on, $5 CAD a month.
Build your own interface.
Organizations already run their own branded interfaces on Emily, such as a client's custom coaching app. You design the experience; Robson and Emily do the rest.
You build
Your interface
The screens, branding and workflow your people see.
Robson runs
Sign-in and access
Every request is authenticated by Robson, wherever your interface runs.
Emily provides
Memory and reasoning
Long-term memory, the Epistemic Firewall and your chosen AI model.
Where your interface runs, for example on a developer's laptop while you build or on your own server in production, is agreed when we issue your access. Sign-in always goes through Robson.
From request to live.
Anyone can register with Google and use Emily today, on the web, by email, through MCP or with the SDK. For a custom interface or a white-label build, this is how it works.
Request access
Tell us what you want to build and who will use it.
Agree scope and credentials
We agree how your app will connect, where it will run, and issue your credentials.
Build and test
You build against Emily with the API reference and examples we share with you.
Go live
We confirm your production setup, and your people start using it.
API reference and examples
The endpoint reference, authentication guide and quick-start code are shared with developers when access is issued.
Tell us what you want to build.
Describe your app and the people who will use it, and we will set up developer access.
Or email sales@robsoninc.com