Adafruit IO
Adafruit IO is a cloud platform for storing, visualizing, and acting on data from connected devices and Internet of Things projects.
Verdict
Common use cases
- Pull temperature logs from field sensors
- Push test results to IoT dashboard
- Archive device telemetry before hardware swap
- Compare feed data across prototype iterations
- Audit dashboard configs for client demos
Integration
- Vendor
- Adafruit IO
- Category
- other
- Auth
- API_KEY
- Tools
- 21
- Composio slug
adafruit_io
Tools
- Add Feeds to Group
Sequentially add 1-10 existing feeds to one group using one documented provider request per feed. This is non-atomic: processing stops after the first failed or ambiguous attempt, while earlier successes remain applied and are returned with
- Create Dashboard
Create a dashboard for visualizing and controlling feed data.
- Create Feed
Create a feed, optionally within an existing group.
- Create Group
Create a group for organizing feeds.
- Delete Dashboarddestructive
Permanently delete one dashboard by exact key.
- Delete Data Pointdestructive
Permanently delete one feed data point by ID.
- Delete Data Pointsdestructive
Permanently delete 1-30 feed data points sequentially through Adafruit IO's single-data-point DELETE endpoint, bounded by the Free-plan limit of 30 data mutations per minute. This operation is not atomic: each ID has an independent outcome,
- Delete Feeddestructive
Permanently delete one feed and its retained data by exact feed key.
- Delete Groupdestructive
Permanently delete one group by exact key.
- Get Account Status
Return compact identity, plan limits, and current data-write throttle usage for the connected Adafruit IO account.
- Get Dashboards
List dashboards, or retrieve one dashboard and its blocks by key.
- Get Data Point
Retrieve one feed data point by its exact ID.
- Get Feeds
List feeds, or retrieve one feed by key with optional data summary details.
- Get Groups
List groups with feed summaries, or retrieve one group by key.
- List Feed Data
Return one newest-first page of feed data with an opaque cursor for the next older page.
- Publish Feed Data
Publish one or multiple timestamped values to a feed, using the batch endpoint only when multiple points are supplied.
- Remove Feeds from Groupdestructive
Sequentially remove 1-10 feeds from one group using one documented provider request per feed. This is non-atomic: processing stops after the first failed or ambiguous attempt, while earlier successes remain applied and are returned with the
- Update Dashboard
Update an existing dashboard by exact dashboard key.
- Update Data Point
Replace the value and optional metadata of one existing feed data point.
- Update Feed
Update an existing feed by exact feed key.
- Update Group
Update an existing group by exact group key.
Setup
Setup guide
- 11. Open your Switchy workspace and navigate to Settings > Integrations. 2. Click 'Add Integration' and select Adafruit IO from the list. 3. Log into your Adafruit IO account at io.adafruit.com and go to My Key in the top navigation. 4. Copy your Active Key (the long alphanumeric string under your username). 5. Paste the key into Switchy's API Key field and click Connect. 6. Switchy will verify the key by fetching your account details — you'll see a green checkmark if it succeeds. 7. Open any Project, type '@Adafruit IO list my feeds' in the chat, and confirm you see your existing feeds in the response. 8. If you get an authentication error, regenerate your key in Adafruit IO and repeat step 5.
What teammates see: by default, memories from Adafruit IO are scoped to the Project (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Projects have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
Summarize Recent Sensor Data
@Adafruit IO fetch the last 100 data points from my 'temperature' feed and give me the min, max, and average readings over the past day.
Create Dashboard for Demo
@Adafruit IO create a dashboard called 'Field Test Demo' and add my 'humidity', 'pressure', and 'battery-voltage' feeds to it.
Archive Old Data Points
@Adafruit IO delete all data points from my 'prototype-log' feed that are older than March 1st, 2024.
Organize Feeds by Project
@Adafruit IO create a group called 'Solar Panel Tests' and add my 'voltage-output', 'current-draw', and 'panel-temp' feeds to it.
Push Calibration Results
@Adafruit IO add a data point with value 3.27 to my 'calibration-reference' feed and tag it with today's date.
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@adafruit-io show me the last 24 hours of temperature readings from my greenhouse-temp feed, and tell me if anything looks unusual
Retrieved 47 data points from your greenhouse-temp feed over the past 24 hours. Temperatures ranged from 18.2°C to 24.7°C, with most readings clustering between 20-22°C during daylight hours. One anomaly: a spike to 29.3°C at 14:23 UTC yesterday, which held for approximately 18 minutes before returning to normal range. This could indicate direct sunlight exposure or a sensor calibration issue worth investigating.
This example shows read-heavy feed querying paired with AI analysis. The MCP fetches time-series data points; the AI interprets patterns. Note that Adafruit IO's Free plan retains only 30 days of data and limits retrieval rate, so large historical queries may need pagination or fail if you exceed API rate limits.
@adafruit-io create a new feed called soil-moisture-sensor-2 in my garden-sensors group, then log the current reading of 67% to it
Created feed 'soil-moisture-sensor-2' within your garden-sensors group (feed key: soil-moisture-sensor-2). Successfully logged data point: value 67, timestamp 2025-06-14T10:32:15Z. This feed is now visible in your Adafruit IO dashboard and can accept up to 30 data mutations per minute on the Free plan.
This demonstrates the MCP's write capabilities: creating feeds within groups and posting data points. Be aware this is a destructive action—the feed persists in your Adafruit IO account until explicitly deleted. The Free plan's 30-mutations-per-minute limit applies to all data writes, so high-frequency logging from multiple sensors may require throttling or a paid plan.
@adafruit-io compare power consumption across all my smart-plug feeds this week and create a dashboard showing the top 3 energy hogs
Analyzed 5 feeds in your smart-plugs group over the past 7 days. Top consumers: office-heater (avg 1,847W), kitchen-espresso (avg 1,203W), living-room-ac (avg 891W). Created dashboard 'Energy Consumption - Week 14' with three line-chart blocks visualizing hourly averages for these feeds. Dashboard is live at adafruit.io/dashboards/energy-consumption-week-14 and updates automatically as new data arrives.
This synthesis example combines data retrieval, AI-driven comparison, and dashboard creation. The MCP handles both read (fetching feed data) and write (creating the dashboard) operations. Dashboard creation is permanent—you'll need to manually delete it later if no longer needed. Free-plan users are limited to 5 dashboards total.
Use-case deep-dives
When Adafruit IO beats building your own IoT backend
A three-person hardware startup prototyping a soil sensor needs to visualize moisture readings from five devices without writing backend code. Adafruit IO wins here because the Create Dashboard and Create Feed tools let you stand up a working data pipeline in under an hour—no server provisioning, no database schema. The 30-data-mutations-per-minute limit on the free plan means you can log sensor readings every two minutes per device without hitting throttles. If your prototype graduates to production with hundreds of devices or sub-minute polling, you'll outgrow this MCP and need a dedicated time-series backend. For early-stage hardware validation where speed matters more than scale, Adafruit IO keeps your team focused on the sensor logic instead of infrastructure.
Managing student hardware projects without IT overhead
An engineering instructor running a 12-student embedded systems course needs each team to publish temperature data from their Arduino projects to a shared dashboard. The Create Group and Add Feeds to Group tools let the instructor organize feeds by team without students needing separate accounts or complex permissions. The API key auth means students paste one token into their sketch and start publishing—no OAuth flow, no credential rotation. The Delete Data Points tool handles the inevitable bad data from wiring mistakes or runaway loops. This setup breaks down if you need per-student access control or audit logs; Adafruit IO treats the API key as a single identity. For a semester-long course where simplicity trumps security and you're okay with shared write access, this MCP eliminates the usual IoT onboarding friction.
When hobbyist scale justifies the free-tier constraints
A solo maker tracking energy usage across four smart plugs wants historical charts without paying for a cloud service. Adafruit IO's 21 tools cover the full CRUD lifecycle—create feeds for each plug, log wattage readings, delete outliers from sensor glitches, and build a dashboard to compare weekly trends. The free plan's 30-mutations-per-minute ceiling works fine if you log every 10 seconds across four devices. You'll hit limits if you add motion sensors or door contacts that fire on every event, not on a schedule. The Delete Feed tool makes it easy to reset experiments without orphaning data. If your home automation grows past a dozen devices or you need sub-second resolution, you'll migrate to InfluxDB or Home Assistant's native storage. For weekend projects and single-room monitoring, Adafruit IO delivers a complete stack with zero setup cost.
Frequently asked
What does the Adafruit IO MCP do in Switchy?
It connects your Adafruit IO account so AI agents can read sensor data, create feeds and dashboards, and write data points to your IoT devices. The MCP exposes 21 tools covering feed management, group organization, dashboard creation, and data point operations. Useful if your team runs hardware projects that log temperature, motion, or other sensor readings to Adafruit IO.
Do I need an Adafruit IO API key to connect?
Yes. The MCP uses API key authentication, which you generate from your Adafruit IO account settings. No OAuth flow—just paste the key into Switchy when you add the integration. Anyone with the key has full read-write access to your feeds and dashboards, so treat it like a password and rotate it if a team member leaves.
Can the MCP delete multiple data points at once?
Sort of. The Delete Data Points tool removes 1-30 points sequentially by calling Adafruit IO's single-point DELETE endpoint for each one. It's not atomic, so if one deletion fails midway the earlier ones still happened. Free-tier accounts hit the 30-mutations-per-minute rate limit quickly, so bulk deletions can take multiple minutes or require a paid plan.
Why use this instead of Adafruit IO's web dashboard?
The MCP lets AI agents act on sensor data without you manually clicking through the Adafruit IO UI. An agent can read a temperature feed, decide the value is too high, then create a dashboard alert or write a control signal back to your device—all in one conversation. The web dashboard is still better for visual exploration and one-off edits.
Who on the team should connect the Adafruit IO MCP?
Whoever owns the Adafruit IO account and can generate an API key. Since the key grants full access to all feeds and dashboards, only connect it in Switchy workspaces where every member should see your IoT data. If you're on a Free plan, remember the 30-data-mutations-per-minute limit applies to all API activity, including this MCP.