Agents change, but your memory stays with you

Memories from different agents come together; their tasks become your daily schedule.

Your memory follows you, not your tools

Use one agent today and another tomorrow — your habits, project conventions and past pitfalls are still there. No re-explaining the background at the start of every session, and no pasting the same instructions into several places.

A schedule you don’t have to build

The tasks your agent takes on, when it plans to do them and how far it has got land on your calendar, and the progress updates as it works. No entries to create, no statuses to go back and change.

The tedious part gets done along the way

Recording, sorting, updating and wrapping up used to need time of their own. Now they happen while your agent works, and you only look when you want to.

How to start

Connect once, then keep working as usual.

Create an account
One minute and an email address.
Connection prompt for your agent
Send this prompt to your agent in the project where you want Nuto. Remote MCP is configured in the host’s connector settings; CLI setup requires terminal access.
Keep working as usual
Let the agent recall memories and record progress as you work.

The workflow asks your agent to read context before work and record milestones. Automatic triggering depends on the host and its settings.

Connection prompt for your agent

The guide explains the available channels, authorization and ongoing use.

Connect to Nuto for memory and task tracking. Read the guide: /start.md

Use the service environment in the guide without changing this application’s backend address. Prefer CLI + Skill. Only when the CLI cannot be installed or run, fall back to remote MCP if the server advertises it and this host supports it. For the CLI channel, I authorize checking/installing nuto and installing/loading the Nuto skill in this project while preserving existing instructions. For MCP, use the advertised URL in the host connector settings and browser authorization; do not require a local CLI.

Never ask for passwords or secrets. After approval, read context and verify the granted permissions. Check duplicates before saving a useful project fact and retrieve it if write access is granted; do not create filler memories. Report the channel, results and unfinished steps. Keep the lifecycle guidance in persistent project instructions; MCP alone does not create automatic hooks.

When the CLI needs me to pair, put the pairing code at the end of your reply, highlighted with a Markdown heading and bold, for example:
## 👉 Finish pairing
**Pairing code: `XXXX-XXXX`** (valid for 10 minutes)
Open the URL the CLI printed, or enter this code in the Nuto mobile app, and approve.
Keep the login command running until I approve; if the code expires, run login again for a new one.

Keep checking the authorization result while the host allows it. If my action is required, tell me the next step. When resuming, check the existing connection first; for the CLI, use nuto auth whoami before starting another login.

Want to review the steps yourself?Open the integration guide.

Frequently asked questions

Do I have to replace my current tools?: No. Nuto works alongside the agents you already use, so you keep working the way you do.
Which agents are supported?: Agents that can run terminal commands connect through the CLI + skill; agents that support remote MCP can add it in their connector settings. The integration guide has the steps.
Where is my data?: On the server you connect to — one someone else runs for you, or one you deploy yourself.
What if I want to stop?: Stop any time. Your agents keep working as usual; they just won’t remember last time.

Agents change, but your memory stays with you

Your memory follows you, not your tools

A schedule you don’t have to build

The tedious part gets done along the way

How to start

Create an account

Connection prompt for your agent

Keep working as usual

Connection prompt for your agent

Frequently asked questions