Help & docs

Everything you need to get going.

Set up Dorinda with your AI, learn how delegations, governance, and the Household work, and get answers to the common questions.

FAQ

Frequently asked questions

What is Dorinda?

Dorinda is the accountability layer for the AI you already use. It turns the general AI you already use — Claude or ChatGPT — into an assistant that actually follows through. The core unit is a Delegation: a durable, closed-loop responsibility that is tracked until done, with who owes what, what is waiting, and what is at risk.

Does it use my data to train AI?

No. Dorinda does not train AI models on your data, and it makes no AI calls on the server. Your own AI receives only the data you direct it to over MCP. Connected account data (Gmail/Google Calendar or Outlook/Microsoft Calendar) is used only to provide the feature, never sold, and never used for advertising or model training.

How is it different from a to-do app?

A to-do app stores a list. Dorinda tracks responsibilities to closure: it knows who owes what, what you are waiting on, what is at risk, and it keeps following through — with your own AI doing the reading and drafting, and your approval on anything that matters. The unit is a closed loop, not a checkbox.

Which AIs work with Dorinda?

ChatGPT and Claude today, via a remote MCP connector — you connect the AI you already pay for. Every feature and every doc works the same with either one; there’s no new chatbot to learn. See Using Dorinda with ChatGPT or Using Dorinda with Claude.

Will I actually stick with this?

There’s no new app to open or habit to build — Dorinda works inside the ChatGPT or Claude you’re already in. That’s the whole point: the tools that fail are the ones you have to remember to open.

Is my data private?

Your assistant state belongs to you. The app is a deterministic system of record with no server-side AI. High-risk actions require your approval, connections are least-privilege and revocable, and you can export or delete your data. See the Privacy page for specifics.

How does the Household work?

On the Family plan, up to five members share a Household — a shared calendar, lists, and projects — with roles (Owner, Adult, Teen, Assistant). Every item is scoped shared or private, so the family sees what is shared while each person keeps their own private Dorinda. Personal is single-user; Family unlocks multi-member (up to five).

How does billing work?

Every plan starts with a 14-day free trial — full access, no card required to begin. We only ask for a card if you decide to keep going. After that, Personal and Family bill monthly or annually through Stripe (annual saves ~20% — Personal $144/yr, Family $444/yr). Dorinda never stores your card; Stripe handles payment and taxes. Cancel anytime; see Pricing.

How do approvals work?

You set the rules. High-risk actions — sending an external message, spending money, contacting someone new — always require your approval, right in the chat. Low-risk actions can be auto-allowed. Your AI can prepare and stage an action, but it cannot cross your line on its own.

How do I connect my email?

Connect your Gmail or Outlook (Microsoft) mail and calendar from your Dorinda account settings. Dorinda can then read your inbox to ground follow-ups in what's actually happened, and send email from your own account when you ask — it stages every message for your approval first and never auto-sends. It also reads your calendar for context and can add events. Access is least-privilege and revocable at any time.

Ready when you are.

Create your account, then connect ChatGPT or Claude in under a minute.