Frequently Asked
Questions

SimSense is an MCP server, so any agent that speaks MCP can connect: Claude, Cursor, Claude Code, or one you built. Add https://my.simsense.ai/mcp as a connector or MCP server, connect, and sign in.

In Claude that is Settings → Connectors → Add custom connector; in Claude Desktop it is the + button in the message box, then Connectors → Manage connectors.

No. The first time you connect, SimSense asks you to sign in with Google or GitHub and creates the account for you. You can also sign in first at my.simsense.ai.

Yes. Your account is the same everywhere, so every session sees the same devices, sims, agents and kits.

Any MCP client. Claude is the one we document most closely, but the server is the same for all of them, and the kits and hosted agents work whichever client you use.

Anything that can report what it senses, show a screen, or carry out a command. A spare phone, a tablet or TV with a browser, a camera, a microcontroller such as a Presto or ESP32 board, a Raspberry Pi. A device tells SimSense what it can do when it connects, and agents work from that.

From the dashboard choose Add device. For a phone or tablet you scan a code and it is paired; a phone can then run as a sensor, a screen, or both. Hardware you build registers itself with one request, described in the developer docs.

Try it without an account on the phone sensor page.

Motion, orientation, compass, sound level, light, touch, network and location, and it can take a picture on request. You choose which sensors to enable before granting permission, and you can change them later.

Yes. A device declares the commands it accepts, and an agent, or you through your own agent, can send them: change what a screen shows, push an image, switch an output. Results come back so the agent knows what happened.

Anything with a web browser: tablets, smart TVs, a Fire TV Stick, an old laptop, a Raspberry Pi with a monitor, or a device with a display that SimSense drives directly. A screen shows one sim at a time and you can switch it whenever you like.

An agent is a hosted worker that runs a Claude model against your devices, your sims and any services you connect. You give it a job in plain language. SimSense runs it on a schedule, or wakes it when something happens.

On a schedule you pick, or in response to an event: a device reading, a change on a sim, or a condition you define such as "the fridge has been above 8°C for ten minutes". Your plan sets how many triggers and conditions you can have and how often an agent can wake.

A kit is a device, an agent and a sim with one job, packaged so it installs in one step. Install it, connect the devices it asks for, and it runs. The setup is a conversation, not a form, and you can keep asking for changes afterwards.

Yes. Build it with your agent through the connector, save it privately, then review and publish it on the web. Published kits appear in the community gallery with your creator profile, and other people can install or remix them with attribution.

Claude models, hosted by SimSense. You do not need your own API key. Model usage is metered against your plan's monthly allowance.

A sim is a live web page that shows state: a menu, a board, a poll, a dashboard. Devices and agents write to it and it updates in place on every screen showing it. Most sims arrive as part of a kit; you can also ask Claude to make one.

Ask. A sim installed from a kit has a steward conversation on its page where you describe the change. Or tell your own agent directly through the connector. You can also manage sims from the dashboard.

Each sim and device is private, shared with your organization, or public. A public sim is open to anyone with the link, and you can additionally list it so people can discover it. Organizations can require everything their members make to stay private.

Three devices, three sims and one agent, with a monthly model allowance sized for a first kit to run properly, one trigger and one condition. No card needed.

Pro raises the limits to 25 devices, 25 sims and 5 agents, with a much larger model allowance, ten triggers per agent and ten conditions.

Team is sold to an organization: every member gets generous per-person limits and the whole team shares one pooled model allowance.

Enterprise removes the limits. Prices are shown at checkout in Settings → Billing.

Creating another sim, device or agent asks you to upgrade. Agents that reach the day's model allowance pause until it resets, and the dashboard tells you. Nothing you have made is removed.

From Settings → Billing, which opens the billing portal. A cancelled plan runs to the end of the period you paid for.

An organization owns a verified email domain. Anyone who signs in with that domain joins automatically. Admins manage members, share devices and sims within the organization, and can set a policy that keeps members' work private.

Start a new conversation after adding a connector. If it still does not show, check it is listed in your agent's connector or MCP settings, update the app, and if needed remove and re-add the connector.

Disconnect SimSense in your agent's connector settings and connect again, signing in with the same Google or GitHub account as before. That refreshes the authorization.

Check the device is online on the Devices page, that its browser is open on its device URL, and that a sim is assigned. Asking your agent to list your devices shows which sim each one is showing.

Open the agent to see its last runs and why the last one stopped. The usual causes are a paused agent, a plan allowance that has run out for the day, or a trigger clamped by the plan's wake rate.

Sign in at my.simsense.ai. Settings holds your connector details, billing and profile.

Email privacy@simsense.ai from your account address. We remove the account as described in the Privacy Policy.

Email hello@simsense.ai, or join the Discord.