flojoy.notion.site Open in urlscan Pro
2606:4700:4400::6812:21e8  Public Scan

Form analysis
0 forms found in the DOM

Text Content

Flojoy

Search
Duplicate


Try Notion

Drag image to reposition
Flojoy


Open-source visual programming for testing, measurement, control, and AI.
flojoy.io

Test, measurement, & control
Industry overview
TMC software vs hardware vendors
Trends upending TMC
LabVIEW: The 500 pound of TMC software
Problems with TMC software
Quotes from Recovering LabVIEW users
Envisioning a solution
Wishlist for modern TMC software
Enter Flojoy
Flojoy product
Visual scripting for Python
CTRL Panel
Product Roadmap
Current Team
Jack Parmer
Asif Hasan
Tristan Britt
Jessica Johnson
Angel Chauffray
5 CS and physics undergrad summer interns
Go-to-market (GTM)
Open-core
Content marketing & university STEM
2023 OKRs
Burn rate
Seed investment
Exit outcomes

Test, measurement, & control
Industry overview

The global test & measurement equipment industry is sized at around $30B. It is
expected to reach $40B by 2030.

Test, measurement, & control (TMC) underpins every research and engineering
endeavor in the world. To illustrate, here are a few examples:

Semiconductors: when a printed circuit board or chip undergoes QA at a test
bench, TMC software and hardware perform the test.
Chemicals: When reagents are synthesized in a reactor, TMC hardware and software
assess chemical properties like molecular weight, solubility, LogP, etc.
Automotive: Every part that makes its way into an assembled car undergoes
inspection with TMC software and hardware.
Computer vision: TMC vendors provide microscopes, lenses, and cameras for most
industrial imaging applications.
Academia: when a researcher measures a material’s response to laser pulses, they
are using TMC software and hardware to capture the measurements.

Naturally, TMC is largely unheard of outside of science and engineering.
However, it underlies every high-volume, high-tech, manufacturing process and
experimental science discovery.

(For readers more familiar with software engineering, an apt TMC analogy is unit
testing and continuous integration. Production software is never written without
unit tests and CI. Analogously, production lines for high-tech products are
never built without TMC at every step.)

TMC software vs hardware vendors

Most of the storied names in TMC are industrial scientific hardware companies.
Below is a partial list of such companies (including their specialty and
revenue). Most of these companies provide bare-essentials desktop software
(predominantly Windows) to operate their machines. In other words, software is
far from the core expertise of most TMC vendors.

Specialty
Revenue/yr (millions USD)
Thorlabs
Optics
500
Stanford Research Systems
Laboratory benchtop instruments
<50
Agilent Technologies
Chemicals analysis
6,300
Keysight
Electronics testing
5,400
Thermo Fisher Scientific
Drug development
39,000
PerkinElmer
Chemicals analysis
29,000
Zeiss
Microscopy
7,500
Leica
Microscopy
400
National Instruments (NI)
Test automation
1,400
FLIR
Thermal imaging
2,000
Illumina
Genome sequencing
3,200


Since most hardware vendors provide their own proprietary desktop software to
operate their machines, there are far fewer TMC software vendors:

Specialty
Revenue/yr (millions USD)
NI LabVIEW
No-code visual programming
1,400
MATLAB
Engineering IDE and programming language
1,200


Trends upending TMC

Over the past decade, there has been a proliferation of viable, non-proprietary
TMC options outside of lock-in with the above vendors. Trends enabling this
include:

Python: The rise of this free, scientific programming language has democratized
serialized communication with science and engineering hardware.
GitHub: Today, most scientists who dabble in code have a GitHub account. TMC
software (like LabVIEW and MATLAB) is not well designed for how scientific
software is shared today on GitHub/GitLab.
Raspberry Pi: Low-cost, single-board computers like Raspberry Pi allow
scientists and engineers to program their own, customizable DAQ devices.
USB: Ubiquity of USB protocols for serial communications (vs more obscure RS-232
or GPIB/IEEE 488 protocols and hardware) allow everyday laptops to connect to
most DAQ devices, without the need for expensive, proprietary adapters.
LabJack: Inexpensive DAQ boards, such as LabJack, have vastly decreased the
expectation for how much DAQ hardware should cost (LabJack is 5x less expensive
than NI’s equivalent).
AI/ML: The success of ML models for computer vision and AI is allowing more work
that was previously done with expensive hardware to be done in software.

LabVIEW: The 500 pound of TMC software

National Instruments (”NI”) is a $1B+ revenue, 7,800 employee, public company.
NI’s flagship software, LabVIEW, is the staple software for data acquisition
(”DAQ”), in science and engineering.

In a nutshell, LabVIEW is GUI-based, desktop software that scientists and
engineers use to connect to robotics and instrumentation. LabVIEW’s main selling
point is lowering the barrier to entry for “talking to machines.” With LabVIEW,
engineers and research scientists can operate instrumentation without ever
writing low-level code.

Without LabVIEW, scientists or engineers typically use a programming language
(such as Python, C#, MATLAB, etc). Since most scientists and engineers do not
have this degree of computer science training, LabVIEW’s GUI has become the de
facto way for research scientists and test bench engineers to interface with
benchtop robotics and measurement equipment.

Below is a screenshot of a LabVIEW visual program. LabVIEW programs are made by
clicking and dragging various nodes on a canvas, then interconnecting the nodes
with “wires.” The nodes represent actions, such as “read an analog signal from
this device” or “compute the derivative of this signal.”



LabVIEW visual programs result in a control dashboard that can look something
like the below screenshot. After being developed by a senior engineer, these
control dashboards are often used by technicians and lower-level employees, to
monitor system operation, run routine tests, administer experiments, etc.



Problems with TMC software

Current TMC software (LabVIEW being the leading example), has several outdated
and vexatious drawbacks. LabVIEW’s creator, National Instruments, seems unaware
or unknowing of how to deal with them. Compared to modern software in other
categories, TMC software drawbacks include:

Outdated cost & licensing model: LabVIEW is expensive ($500-$2.8k per seat, per
year). It is also licensed through a 90’s-era workstation model (rather than a
cloud-based SaaS model). A savvy disruptor could be open-core, shifting
monetization from desktop software to cloud-based data storage, analysis, and
control.
Closed-source: NI has not embraced (or grasped) the open-source/open-core sea
change that is permeating science and engineering software. Despite being a
7,800 employee company, NI OSS projects have 10x less GitHub stars and 10x less
GitHub contributors than Plotly or Databricks.
Lack of integrated analysis: LabVIEW is great for extracting data from machines
and saving it locally, but then relies on other software (such as MATLAB,
Origin, RStudio, Jupyter, etc) to visualize and analyze that data. LabVIEW is
missing a huge opportunity here. A disruptor could stream the extracted R&D data
to a secure cloud that includes an integrated analysis environment.
Poor design: LabVIEW has a 90s-era Windows design with little attention to UI
and UX. A savvy disruptor with a minimalist, clean, and modern design (that is
still familiar to converted LabVIEW users), would underline LabVIEW’s antiquity.
Platformless: Since LabVIEW is desktop software, there is not a high-quality
platform for searching and sharing LabVIEW recipes or interacting with other
LabVIEW users. It should be rather simple to search and download recipes such as
“capture a microscope image every second” or “run a current-voltage curve on a
solar cell.” Manufacturers of scientific equipment could upload recipes for
their hardware to this platform.
Learning curve: Speaking from experience, LabVIEW is not simple to learn. The
proprietary nature of the software does not help the learning curve, given the
lack of online blogs and community docs that guide you through the first-time
user experience. For an engineer that is an expert in Python or MATLAB, it is
vastly easier to interface with hardware in these languages. However, most
scientists are not sufficiently proficient in any programming language,
therefore, LabVIEW’s drag-and-drop UI reigns supreme. Most LabVIEW programs are
created by tweaking legacy programs that have existed for years in the company
or lab. With an open-core business model, along with a social platform where
recipes can be shared, Flojoy will have a delightful, rapid learning curve.
Flojoy app can be installed, modified, and run in under 5 minutes from the time
of initial download.
Runaway complexity: The wires in the visual programming paradigm are great for
simple apps, but can become a rat's nest for complex apps. For example, here is
a screenshot of a dizzyingly complex LabVIEW program:

Speaking from experience, these complex LabVIEW apps are time-consuming to debug
and maintain. LabVIEW apps cannot be committed to source control (like
git/GitHub), so there is a high risk of regression when multiple engineers are
contributing to the same app. Flojoy’s app file format is text-based and
machine-readable (easily compatible with source control). Flojoy’s open
architecture (see below) allows large swaths of unsightly visual programming
wires to be abstracted into a single custom Python script/node. In other words,
Flojoy’s pluggable Python backend allows the end user a customizable level of
abstraction.

Quotes from Recovering LabVIEW users

"I’m very excited by the idea of an open-sourced improvement to LabView. In
principle, it can be very diverse in its applications as well as being well
suited to handle complicated logic, but its frustrating quirks make it very
ill-suited for the inexperienced / uninitiated."
PhD student at Mcgill, runs undergrad physics labs which all use LabVIEW

"I'm a frequent user of Plotly today, and in a couple former lives I was a
LabView user... LabView and their National Instruments boxes are the root cause
of a test stand failure that came very close to killing me - part of the USB
port on the National Instruments chassis failed, sent a bunch of rogue commands,
and pressurized a tank I was standing next to up to 20,000 psi. That tank was
rated for 10,000. And in addition to that, I still have tendonitis in my left
forearm that started when I was doing significant development in LabView and hit
"Ctrl+E" to flip between the interface and block diagram way too many
times. That said, LabView is a terrible piece of software and I'm glad you're
going after it."
Frmr CTO of a major aerospace startup

"I am really curious about your new initiative. It turns out that "open-source
LabVIEW" (in one form or another) has been floated around in discussions with
our clients as something that we wished existed."
CEO of scientific Python consulting firm

Envisioning a solution
Wishlist for modern TMC software
After over a decade of using LabVIEW, speaking with LabVIEW users, and writing
TMC software, my dream TMC software would be:
No code with a backdoor for coders. I want to whip together quick DAQ
applications, but also to be able to fine-tune nitty gritty detail in code.
Python-based. Python has great DAQ foundations with libraries like PyVISA and
PySerial. However, these projects need funding, love, and increased attention,
in order to rival the billion-dollar companies listed in the first section. They
also need a visual no-code interface to lessen the barrier to entry.
Open-core/open-source. The best scientific software being published today is
open-source. Open-source allows large communities of support to grow around the
software. Flojoy is launching with an open-core model similar to RStudio.
Slick UI. Creating visual programs that interface with the physical world should
be creative and joyful. I want a clean, no clutter, and drag-drop interface that
allows real-time exploration of the physical world.

Enter Flojoy

Flojoy was born from the above wishlist. With Flojoy, we are building a free,
open-source, and GUI-based desktop program for modern TMC. The software
architecture and ethos represent an evolution from LabVIEW in several ways which
have already been discussed. Specifically, Flojoy is:
Open-source: Flojoy is free to download and licensed under the AGPL license.
Web-based: Rather than written in a desktop language such as Java or C++ (like
LabVIEW), Flojoy is written primarily in Typescript, with Electron as the
desktop packager (same architecture as VSCode). This architecture enables a more
elevated UI/UX design, access to higher quality software talent, utilization of
OSS web libraries (such as Plotly.JS & React Flowchart), mobile support, and
online portability.
Open-architecture on GitHub: Each node in a Flojoy program represents an
open-source Python script that is managed and supported publicly through an
official Flojoy GitHub repository. These nodes/Python scripts can be contributed
to GitHub by any community member or hardware vendor, through simple pull
requests.
No-code with a Python backdoor: Crucially, Flojoy users would never need to be
versed in Python, in order to string these scripts together in the GUI (or even
know what Python is!). However, with this open architecture, advanced Flojoy
users could easily add custom nodes to their GUI canvas, simply by writing a
Python script that adheres to the Flojoy API.
Cloud-connected: Data extracted from Flojoy is streamed to flojoy.cloud - a
“research cloud” where it can be viewed, downloaded, visualized, shared,
annotated, and analyzed.
Flojoy product
Visual scripting for Python
Below is an animation of an example Flojoy application (slowed down for demo
purposes).


Each node in this flowchart represents a Python script. When a node lights up,
the Python script (that the node represents) is being executed at that very
moment. In this way, Flojoy applications are built and executed visually by
wiring interchangeable nodes together in a logical order. This makes it easy to
rapidly build Python-based data pipelines in Flojoy, even if you do not know how
to program in Python.

As of January 2023, Flojoy ships with a few dozen nodes (Python scripts). These
nodes can be mixed and matched in any Flojoy program. They are mostly basic
vector arithmetic and visualization operations. By mid-2023, there will be 100s
plug-and-play Python nodes that ship with the Flojoy software. These nodes will
encompass a multitude of Python engineering operations: data visualization, data
extraction from scientific instruments, cloud ETL, numeric vector and matrix
operations, running ML models, controlling peripheral hardware, and more.
Crucially, Flojoy has been designed with an open architecture, so that it is
very easy to write and contribute custom Flojoy nodes. For example, a community
contributor might want to contribute a set of nodes that read and write from a
particular cloud database that they use. Similarly, a scientific instrumentation
manufacturer may want to contribute a set of nodes to interface with their
product.

Here is a simple Flojoy application that generates a sine wave, adds random
noise to the sine wave, then visualizes the result in 2 charts. Again, each of
these nodes is an editable Python script under the hood.



CTRL Panel

In addition to a canvas for wiring nodes together to build Flojoy programs (the
“IDE”), Flojoy has a “CTRL Panel” for building dashboards and interactively
setting the program’s input parameters. For example, the below control panel
has:
An input knob that allows resetting a sine wave’s input frequency
An output plot that shows the resulting sine wave
When the input knob is turned, the Flojoy program automatically reruns, and the
sine wave output is displayed.

The CTRL Panel is important because an engineer might write a Flojoy program for
someone else to operate. Imagine a Flojoy program that instructs a stepper motor
to move an x-y stage to various positions, while taking photographs of a sample
at each position. An engineer might write this program for a technician to
operate through the control panel. Instead of viewing the entire program as a
graph of nodes, control panel operators can focus on only the crucial inputs and
outputs.

Below is a screencast that demonstrates everything working altogether. This
simple app loads a Flojoy program that counts to 10. The counter digits are
displayed in the CTRL Panel.

Product Roadmap
In 2023, the product roadmap is focused on releasing world-class open-source
software, creating buzz, adding support for popular hardware instrumentation,
and becoming the de facto visual scripting software for Python data pipelines.
Stabilize the software and open-source Flojoy under a copy-left (AGPL) license
Open-source and document over 100 nodes and example programs for the Flojoy user
community
Launch a Flojoy Discourse forum to grow and nurture the community
Develop an online demo site where users can try the Flojoy UI without
downloading the software
Leverage Electron for command-line free installation
Implement telemetry and active monthly user analytics with Mixpanel and Sentry
Demonstrate running Flojoy “on the edge” - directly on single-board computer
devices like the Raspberry Pi. (The project Node Red has demonstrated this
nicely.)
Current Team

The Flojoy team has a unique mix of software and start-up expertise, plus prior
working history with LabVIEW and TMC. We are well-positioned to build the de
facto visual scripting software for Python and disrupt the TMC software
industry.


Jack Parmer
Role
Founder/CEO
Previous
Plotly (8 years as CEO), Domino Data Labs
Education
BS Engineering Physics, Stanford University
Website
jackparmer.com


Asif Hasan
Role
Engr. Director
Previous
Software Engineer II at Amazon/AWS, Toptal
Education
MS Comp Sci, Vanderbilt University
Website
Toptal profile


Tristan Britt
Role
Product Manager
Previous
LabVIEW instructor for McGill undergrads
Education
PhD Physics, McGill University
Website
https://tbritt.xyz/


Jessica Johnson
Role
Project Manager
Previous
Product Manager at Mentor.ly
Education
MS Chemical Engineering, U. Ontario
Website
https://jalium.github.io/online-cv/


Angel Chauffray
Role
Content developer
Previous
Founder of Nefilatek, a 3d printing materials company
Education
MS Applied Physics, EPFL


5 CS and physics undergrad summer interns
Role
Full-stack developers & community support
Previous
CAE, Google, Ubisoft
Education
BS CS and/or Physics at Mcgill (ongoing)

Go-to-market (GTM)
Open-core
Like Plotly, Redis, Databricks, or MongoDB, Flojoy is an open-core product and
company.
For open-core companies, the primary GTM motion is through the distribution of
the open-source software. Along with establishing the open-source software as an
essential standard in the field where it is commercialization (in our case,
TCM/DAQ).
From their experience at Plotly, the team has world-class, repeatable experience
in building open-core companies and products.
Content marketing & university STEM
In addition to marketing through open-source, Flojoy will implement these GTM
strategies in 2023:
University student ambassador program. Companies like SAS, MATLAB, and National
Instruments built their empires by indoctrinating engineering students with
their software at the undergraduate and graduate levels. Giving software away
for free, partnering with flagship labs, holding hackathon contests, and giving
away SWAG are some of the infiltration techniques that these companies employed.
SEO-driven content-marketing. There is not much great educational content online
for connecting to the most popular scientific instrumentation with Python. We’ll
fill this whitespace with both videos (YouTube) and written (documentation
tutorials). This content will be highly SEO-optimized - aimed at winning these
niche yet highly valuable searches.
Syndication with hardware and IoT blogs. The blogosphere for edge devices like
Raspberry Pi is very active. It will be easy to generate backlinks by
syndicating blog posts with the most popular of these outfits.
TikTok learning videos
2023 OKRs
Execute the product roadmap (see above)
Exceed over 1M open-source downloads by 2024
Achieve cash-flow breakeven (combined PS and product revenue)
Establish recurring product revenue and commercial product-market fit.
Demonstrate a path to $1M in ARR in 2024.

Burn rate

Flojoy’s cash burn rate is currently less than $20k/month. We do not expect to
exceed $20k/month in 2023.

At this burn rate, through a combination of professional services and product
sales, we anticipate cash-flow breakeven by 2024. At Plotly, much of the
open-source software was funded through high-margin professional services
contracts - we can repeat this playbook with Flojoy.

Seed investment

Flojoy is currently raising 1M USD in seed contributions through SAFE notes with
a $10M cap.

This funding will be directed at executing the 2023 OKRs listed above.
Specifically,
Developing online content-marketing, video tutorials, and co-marketing
partnerships
Partnering with world-class Python content-creators on open-source content and
syndication
Delivering professional services and Enterprise VPC software to clients
Hiring an ex-LabVIEW sales engineer or AE to develop ARR and PS opportunities
from inbound leads

You will be part of building Python’s first open-source, no-code visual
scripting IDE. Like Plotly, we anticipate Flojoy open-source distribution to
become a standard-bearer - eventually exceeding 100s of millions of downloads.
This is an opportunity to advance a new generation of testing, measurement, and
control software.

Exit outcomes

Flojoy is in a highly-technical space at the interface of hardware and software.
Like Plotly, Flojoy’s users are mainly trained engineers and scientists. Flojoy
currently has no other startup competition in the TMC space.

Eventually, Flojoy will most likely sell to one of the large scientific hardware
vendors that competes with National Instruments (NI) and want to bolster its
software acumen and sales. Among the scientific instrumentation leaders, only NI
has a strong software product (LabVIEW). To invite acquisition opportunities,
Flojoy will develop modules and content for each vendor’s most popular
instrumentation (see Go-To-Market section).

Since Flojoy can run “on the edge” on Raspberry Pi and single-board computers,
clouds making plays in industrial IoT may also be interested in acquiring Flojoy
(eg Azure, AWS, or GCP).

One of Flojoy’s 2023 OKRs (see OKR section) is to demonstrate a path to $1M in
product ARR in 2024, which would value Flojoy at $10M-15M after its 2nd year of
operation.