Using Notehub IQ
In this tutorial you'll use Notehub IQ, the intelligence layer for Notehub, to explore your project’s data and see how AI can help you understand and analyze it. Then, you’ll learn how to connect your own AI tools to Notehub IQ to build powerful automations.
Introduction
This tutorial should take approximately 15-20 minutes to complete.
In this tutorial, you'll learn how to:
- Connect to Notehub IQ in the browser, with nothing to install.
- Onboard a Notehub project to Notehub IQ, and create datasets from the data your product has been sending.
- Ask questions about your project data.
- Connect Notehub IQ to your own AI agent, and have it check on your product on a schedule.
This tutorial assumes you have an existing Notehub project with data. If not, complete the Writing Host Firmware tutorial before continuing.
Notehub IQ is currently in public beta, and it's free to use during the beta. If you run into issues or have feature requests, send us your feedback.
What Is Notehub IQ
Notehub IQ is an intelligence layer for Notehub. It lets you ask questions about your product in plain language and get answers drawn from the data your Notecards already send to Notehub, without building dashboards, exporting data, or writing queries.
Under the hood, Notehub IQ is made up of two pieces that Blues runs for you:
- A first-party storage service that securely stores your project's events and organizes the Notefiles you choose into queryable datasets.
- An AI interface, the Notehub MCP, that turns your questions into queries, runs them against those datasets, and reasons over the results.
You can access Notehub IQ in two ways:
- In your browser: Blues AI, built into this Developer Center, can connect to your Notehub projects. There's nothing to install, making it the fastest way to try Notehub IQ. You’ll use this approach for most of this tutorial.
- With your own AI agent. Notehub IQ is also available through the Notehub MCP, a hosted Model Context Protocol server that works with MCP-capable clients such as Grok Bot, Claude, Claude Code, Codex, Cursor, or GitHub Copilot. Using your own agent lets you combine Notehub IQ with the other tools your agent can access and monitor your product on a schedule. You'll set this up at the end of this tutorial.
Both options connect to the same Notehub IQ service, so the projects you onboard and datasets you create are available regardless of which AI agent you use.
Notehub IQ can make changes to your Notehub projects, but Blues AI always asks for your approval first.
Connect to Notehub IQ
Now that you have some background on what Notehub IQ is, let’s try it out by connecting to it in your browser.
-
Open Blues AI, which appears in one of two places in your browser window:
-
A message box at the bottom of your browser window. Click into the Ask Blues AI a question… box to open it.
-
A tab on the right-hand side of your browser window. Click the sparkle icon to open it.
-
-
Within the sidebar that opens, click Connect to Notehub IQ.
-
If you're not already signed in to Notehub through the Blues Developer Center, sign in when prompted.
-
Next, authorize the Blues AI assistant to access your Notehub account on your behalf. You can optionally use the Limit Access To dropdown to restrict the assistant to a single project, but don't pick a Viewer Privileges option, since you'll need to make changes when you onboard your project in the next step. You can suspend or delete this access at any time from Notehub’s API Access page. When you’re ready, click Authorize to continue.
-
If all went well, Notehub IQ should now show as Connected within Blues AI.
Now that you’re connected to Notehub IQ in Blues AI, you’re ready to onboard your first project.
Onboard Your Project
Onboarding a project enables an LLM-optimized storage service for that project's data. Before you can use Notehub IQ with a project, you must first onboard it.
To onboard a project you must be an Owner or Developer of the project. If you created the Notehub project yourself, you're already its Owner. Once a project is onboarded, anyone with access to the project, including Viewers, can ask questions about it.
-
Ask Blues AI to onboard the project you wish to interact with.
PromptOnboard my [project name] project.
-
Blues AI will ask you to confirm the changes before proceeding. Click Show details to review the changes, then click Confirm to continue.
After onboarding your project, Notehub IQ prompts you to create one or more datasets. A dataset is a queryable representation of your data that, in most cases, mirrors the contents of your project's Notefiles.
-
Start by creating a dataset for your project's primary data Notefile. If you've completed the Writing Host Firmware tutorial, this will be
sensors.qo, which contains temperature and humidity data. The prompt below also asks for signal quality, so you can check your devices' cellular connections later.PromptCreate a dataset for my sensors.qo Notefile, and include the signal quality fields (bars, RSSI, RSRP, RSRQ, and SINR) from each reading's session.
If your project stores its data in a different Notefile, use that Notefile's name in the prompt instead of
sensors.qo.note Because LLMs are non-deterministic, the agent may present you with different questions about how your data is stored. For example, it might ask whether the temperature values are in degrees Celsius (they are), whether the humidity values represent relative humidity as a percentage (they do), or whether you’d like to rename any fields (up to you). It may also propose more than one dataset for a single Notefile.
-
Next, create two datasets from
_session.qo, the System Notefile where Notehub records the start and end of each device session:firmware, which tracks which Notecard firmware version each device is running.session_usage, which tracks how much data each session sent and received. It only includes session-end events.
PromptCreate the firmware and session_usage datasets for my project.
Notehub IQ may have already proposed these datasets alongside your first one. If so, confirm them and move on to the next step.
-
Finally, create a dataset for
_health.qo, a System Notefile that records device health events such as Notecard restarts.PromptCreate a dataset for my _health.qo Notefile.
-
Click Confirm when Blues AI asks you to approve each dataset creation.
-
After each dataset is created, Notehub begins copying your existing events into it. This process is nearly instantaneous for new projects, but can take many minutes for projects with large amounts of historical data. Once the initial copy is complete, new events are automatically added to your datasets as they arrive in Notehub. Before continuing, make sure your datasets are ready by asking Blues AI:
PromptAre my datasets ready yet?
With your project onboarded and your datasets ready to go, you’re now ready to ask your first questions.
Ask Your First Questions
With your datasets ready, it's time to start exploring your data with Notehub IQ. You can ask whatever questions you’d like, but here are a few examples to help you get started and see what’s possible.
What’s the latest temperature and humidity reading from my device?
Show my device's temperature readings over the last 24 hours as a line chart.
How much data has my device sent and received today?
How good is my device’s cellular connection?
Which Notecard firmware version are my devices running?
These examples are just a starting point. As your project grows and collects more data, you can use Notehub IQ to explore trends, identify anomalies, and analyze behavior across devices. For example:
Do you see any concerning outliers in my device data?
Are any of my devices using an unusual amount of data?
Have any of my devices been restarting unexpectedly?
To get even more out of Notehub IQ, you can connect it to your own AI agent, where you can automate recurring checks and integrate with other tools and services you or your organization uses. Let's look at how to do that.
Watch Your Product With Your Own AI Agent
Blues AI is the quickest way to start talking to your product, but it only responds when you ask. By connecting Notehub IQ to your own AI agent, you can automate recurring checks, integrate with other tools and services, and have your agent proactively monitor your product and alert you when something needs attention.
You can connect Notehub IQ to any AI agent that supports the Model Context Protocol (MCP), including Grok Bot, Claude, Claude Code, Codex, Cursor, and GitHub Copilot. Each connects through the same hosted Notehub MCP server, though the setup process varies by client.
The first time your agent connects, you'll need to sign in to Notehub and authorize access, just as you did with Blues AI earlier in this tutorial. The projects you've onboarded and datasets you've created are already available, so there's no need to set them up again.
For this tutorial, we'll use Grok Bot, but you can find instructions for other clients in the Notehub MCP documentation.
Getting Started With Grok Bot
In this section you’ll set up some automated checks of your product data with Grok Bot, xAI’s assistant.
We recommend Grok Bot because it’s a cloud agent: it runs on its own always-on computer rather than on your laptop, so it can check on your product on a schedule, reach the tools you already use like Slack and email, and message you when something changes, even while you’re asleep. It also simplifies connecting to other services and managing credentials, so there’s very little configuration required.
-
Download and install Grok Bot. Grok Bot is available for Windows, macOS, Linux, as well as iOS and Android.
-
Follow the prompt to sign in and create an account. Grok Bot offers free trials that are usage based, but you will have to enter a credit card as part of the sign-up process.
tip To check how much of your free trial remains, click your profile picture in the bottom-left corner of Grok Bot and look for Trial usage.

-
Once you’re signed in, ask Grok Bot to connect to Notehub MCP. You will be prompted to authorize Grok Bot to access your Notehub account as part of the process.
PromptAdd a custom MCP server named Notehub using https://mcp.notehub.io/mcp.
-
Next, ask a basic question to ensure Grok Bot can retrieve data from your Notehub project:
PromptShow me the latest reading from my [Your Project] Notehub project.
-
With Notehub data being pulled in successfully, you’re ready to give Grok Bot a recurring task that takes advantage of Grok Bot’s cloud-based execution. For example:
PromptEvery weekday morning at 8am, send me a summary of the device activity on my project. Highlight problems such as missing sensor readings, abnormal data, or device reboots.
Review the output and ensure Grok Bot created the routine as expected, and then add:
PromptRun the routine once now so I can see what it looks like.
-
(Optional) Connect apps so Grok Bot is able to take actions using other services you or your company use. For example:
-
Connect either Gmail or Outlook, and have Grok Bot send you a daily email instead of a daily notification.
-
Connect Atlassian, and have Grok Bot log its daily reports to Confluence, or create Jira tickets when it detects device issues for triage.
-
Connect GitHub, and have Grok Bot analyze your device data for potential firmware bugs, then propose fixes as pull requests for your team to review.
-
Connect Microsoft Teams, Slack, or Twilio, as alternative platforms for sending notifications to yourself or your team.
-
What's Next?
You've now connected Grok Bot to Notehub IQ, used it to explore your device data, and set up an automated routine to monitor your product. You've also seen how connecting other apps can turn those insights into actions.
But this is just the beginning. Think about the tasks you or your team perform regularly: checking device health, investigating unusual readings, tracking connectivity issues, or compiling reports. What if an AI agent could handle those tasks for you, automatically?
Experiment with different prompts, schedules, and integrations. Start with something simple, see what works, and build from there. With Notehub IQ providing access to your device data and AI agents connecting that data to the rest of your tools, there's a lot you can automate—and plenty of possibilities we haven't imagined yet.
Next Steps
Congratulations! You've connected Notehub IQ to your project and asked your product its first questions. If you're following a Blues Quickstart, you're done!
Now that you know the basics of how Notecard and Notehub work, you're ready to apply what you've learned towards a real project. We have a few recommendations on what to do next:
-
Send Data to Your Cloud: Notehub IQ answers questions about your data, but many products also need that data in a cloud platform or a system of their own. Notehub routes let you securely forward your events to AWS, Azure, Google Cloud, MQTT brokers, IoT dashboards, and any HTTP endpoint. This hands-on tutorial walks you through creating your first route.
-
Get Your
hub.setConfiguration Right: Yourhub.setsettings determine how and when your Notecard connects and syncs with Notehub, so it's important to tune those settings to your project's power and data needs. For a deeper look, watch An In-Depth Guide to Notecard's hub.set Request, which steps through real-world configuration scenarios. -
Learn Best Practices for Production Projects: This skimmable guide has a list of best practices for moving your project from a prototype to a real, production-ready deployment.
-
Browse Example Apps: Our collection of IoT apps that show how to build common projects with Blues. These example apps include dozens of official Blues project accelerators and a collection of purpose-built sample apps.
-
Follow the Notehub Walkthrough: Learn how to manage projects and device fleets, update Notecard and host firmware, set up alerts, securely route your data, and more.
At any time, if you find yourself stuck, please reach out on the community forum.