Connect · Automate · Secure
Wire up devices and APIs, build visual workflows with AI, and scan every file upload for threats, all in one platform. Real products ship in minutes, not months.
Tendrl
from tendrl import Client
client = Client()
client.start()
client.publish({"temp": 22.4, "humidity": 61})
# validated, routed, and stored.The problem
Most projects stall at the integration layer. You end up managing brokers, writing webhook handlers, and debugging silent failures instead of building your actual product.
| Doing it yourself | With Tendrl | |
|---|---|---|
| Connecting devices | Set up an MQTT broker, manage certs | Install SDK, call .publish() |
| Data quality | Write validation in every service | Declare rules once, enforced at ingress |
| Integrations | Glue code for each API | Drag connectors into a workflow |
| When things break | Check 4 dashboards, grep logs | One trace shows the full picture |
| Adding AI | Another service, another API key | Drop an AI node into your workflow |
Three layers. One platform.
Contact handles connectivity and data. Strand handles what happens next. Surface handles what could go wrong.
Register devices, send data over MQTT or REST, and enforce validation rules before anything hits your system.
Build workflows visually: connect triggers to actions with logic, AI, and cloud connectors. No code required.
Send a file, get a verdict. ML-driven detection catches what signatures miss, without a security team or a signature database.
No toolchain required
Plug an ESP32 into USB and open the Device tab. Tendrl flashes
MicroPython, writes the Wi-Fi and API key, installs the SDK, and
hands you a live terminal and file editor, with no
esptool, no drivers, no serial monitor, and nothing
to install.
.binIt runs on WebSerial, so it is genuinely local: your API key and Wi-Fi password travel from the form down the USB cable, never through our servers.
See the device workbench
Flash, provision, edit files, run scripts, and preview the camera, all from a browser tab
See it in action
Flash a board, provision it, edit files, and run scripts from a browser tab
Prebuilt detectors running on the board: tune by eye, then export the deployable file
Drag-and-drop nodes, connect triggers to actions, add branching and AI
Full execution traces, step-by-step debugging, performance metrics
Developer experience
Connect over MQTT or REST. Every SDK handles connection management, batching, retries, and offline storage, so you write application logic, not infrastructure code. The complexity is abstracted. The performance isn't.
Ideal for Raspberry Pi, servers, and data pipelines. Auto-batching and SQLite offline cache.
uv add git+https://github.com/tendrl-inc-labs/python-sdk 
For high-throughput services. Thread-safe with BoltDB persistence and auto heartbeats.
go get github.com/tendrl-inc-labs/go-sdk 
Browser and Node.js. React hooks, IndexedDB offline storage, and batching built in.
npm install github:tendrl-inc-labs/js-sdk 
Built for ESP32, Pico W, and OpenMV cameras. MQTT over TLS, JPEG streaming, and on-device vision.
mpremote mip install https://app.tendrl.com/api/public/sdk/v1/latest/mpy/package.json 
Classify, summarize, or generate, with memory and tool use
Give AI nodes tools via MCP, and plug Contact, Strand & Surface into any LLM to build faster
Intelligence layer
Drop an AI node into any workflow. It gets conversation memory, configurable temperature, and automatic retries. Or expose your entire platform to AI assistants via MCP.
Monitor temperature, humidity, and motion from ESP32s and Picos. Get alerts in Slack when thresholds cross. Store history in S3.
Learn more →Receive a webhook, validate the payload, enrich it with AI, route through approvals, and push to any downstream service.
Learn more →Classify support tickets, summarize documents, and generate reports, all as workflow steps alongside HTTP calls and database writes.
Learn more →Flash a bare ESP32, provision it, and run code on it from a browser tab. No esptool, no drivers, nothing installed.
Learn more →Run a prebuilt person, face, or hand detector on an OpenMV camera and publish the answer, not the video.
Learn more →Scan what users type, not just what they upload. Stop prompt injection before it reaches your model.
Learn more →Simple enough to start in minutes. Engineered to run in production. Security, performance, and observability are built into every layer, not bolted on after.
Get started
Create an account and your first project
Install an SDK or send data via REST
Add rules so only clean data gets through
Build a workflow with triggers and actions
Go live, monitor, and iterate
Free tier includes 5 entities, 10 workflows, and 550 monthly runs. No credit card. No sales call. Just start building.