Group 2 Developer Tools and 3D Design - Cohort #4 - Workers Launchpad Demo Day
Originally aired on March 6 @ 1:00 PM - 1:30 PM CST
Join us for Workers Launchpad Demo Day, where you’ll have a chance to meet all of the startups selected for our annual cohorts.
Companies:
If you’re a founder building on Cloudflare Workers, head to cfl.re/launchpad to learn more about the program and apply to be a part of Cohort #5 -- applications are now open !
If you’re an investor or a potential customer looking to get in touch with any of the presenting companies, please reach out to the Workers Launchpad team.
Watch the full Workers Launchpad Demo Day for Cohort #4 here !
English
Transcript (Beta)
Hi, my name is Amadeo and I'm the founder of OpenInt. OpenInt helps developers ship native product integrations and enable their customers to connect their apps.
As developers, we know how hard it is to scale product integrations.
This is because every app you want to connect to has a slightly different integration and onboarding process.
It can take weeks to get approved to connect to certain apps and the APIs are often poorly documented.
Existing unified API offerings just don't make the cut because they are too generic to be able to power deep integrations.
And it's really simple to get started.
All you have to do is go to your console and select which integrations you'd like to enable.
You can then launch it without having to wait for any approvals in the app marketplaces.
And you can then share the link with your customers manually or simply embedded within your product to get a full-blown integrations experience.
And here's how it looks like.
We have an example of an app called Acme that added a Greenhouse and QuickBooks integration by just embedding a simple React component.
Think of it like a Stripe subscription or checkout portal but just focused on integrations.
But here's the thing. We think that this is a massive market. Over 40% of both B2B and B2C software has integrations.
And this number only rises with AI agents, which are apps that automate workflows and also need deep native integrations.
To bring it all together, we are OpenIt and we help developers ship native product integrations.
This brings the time to launch integrations from weeks to hours without any approvals.
OpenIt is right at the center of the AI revolution with our open source approach.
And we are well positioned to capture this growth.
You can check us out at openit.dev or just drop me an email.
Thank you. Hey, this is Nis, founder and CEO at Fibroplane, and we're building API developer portals specifically for the HANA TypeScript API framework and the Cloud Flash Stack.
So in the age of AI, everything is an API. And on the one hand, we've got API owners that have several challenges, including API observability, API analytics, API debugging, as well as a slow feedback loop with their users when they want to achieve co-development and time to first integration to be very fast.
For API users, they want quick onboarding, they want great documentation and example workflows on how to get started with an API.
So that's why we've created the Fibroplane API developer portal for these two personas.
More to come in the demo, but basically it consists of a great collaboration and feedback loop for API owners, as well as API analytics and observability based off OpenTelemetry, an AI assistant panel that helps you debug errors that come in, and an authentication service that we're rolling out soon.
At the same time, for API users, we have an API playground for them to test and explore an API.
We can map endpoints to use cases and workflows.
There's great documentation based off OpenAPI, and they can share feedback and issues through the built-in feedback mechanism.
In terms of market, you know, the edge computing market is going to reach $165 billion by 2030.
That's a wave that we're riding. Cloudflare's own growth year-on-year is more than 25%, adding 100,000 developers per month.
That's also a market that we're addressing, and there's over 25 million developers using API developer tools right now.
And of course, APIs are foundational for AI integrations, and there are more developers building these every day.
The team. We've built various developer tool companies in the past.
We're an experienced team in that arena, and we found these companies have been acquired both by Oracle as well as IBM, and our investors are one of the greatest out there.
We've got Target Global, North Zone, Notion, Crane, and a great set of angels.
For the purpose of this demo, I am the API owner, and my buddy Lau is an API user that is exploring my API.
In order to get started with the Fiberplane API developer portal, I've installed Fiberplane Insults and imported the middleware package.
I've mounted that middleware on the SlashFP endpoint inside of my Hano API, and now my portal is available and ready for Lau to be tested and explored.
Hey, I'm Lau, and I'm the API user in this scenario.
So I've opened this playground on the SlashFP route, and in it I can do anything that I want to understand this, how it works.
So I can read the docs about it, I can send requests and get responses, but sometimes you run into issues using the API.
So let's say there's this post endpoint that I fire off, I send requests with just an empty body, and I get this error.
So what I can do is I can report this issue back to the API owner so that they can see that I've run into a 500, which is something that they didn't handle on their end, and so I want them to be able to be aware of this.
So I can say I just sent an empty request and submit feedback.
The feedback is now received, and the API owner knows what's going on.
All right, back to Mies. All right, so I just got a notification inside of Slack that an error was reported by Lau, so let me go ahead and click on that link, and now we can see the entire open telemetry trace of that request and see what's going on.
You can also see how long it took for that request to process, and we get this AI-generated summary of what seems to be the problem centering around a JSON input that was malformed.
So that's Fireplane in a nutshell. There's more to come.
We're going to add more things around API observability and analytics as well as improvements to the AI Assistant.
Stay tuned. Hi, I'm Natalie, and I'm a co -founder of GitLib.
GitLib is an all-in -one coding platform that integrates collaborative coding with Git repositories and one-click deployments.
Our goal is to empower anyone to build software powered by AI with the help of AI.
Every developer has more ideas than time to build them.
The problem is that the process of creating software is too involved, even for experienced developers.
This holds increasingly more for non-developers, and with the advent of AI, anyone has an opportunity to realize a greater fraction of their ideas.
Everyone should be able to produce software to solve their unique challenges or automate toil by easily creating integrations, extensions, bots, or websites.
Our primary target customers are developers who want a streamlined development experience, zero lock-in, and to quickly implement their ideas using the best-in-class development techniques like version control, peer programming, built -in CI, serverless hosting, and human and AI collaborative coding.
While our product is targeted towards experienced developers, we aim to expand the market by making the industry gold standards approachable to everyone.
GitLib is built with our bespoke scalable infrastructure for Git repos and collaborative coding powered by the Cloudflare Workers Platform.
In terms of the opportunity we're pursuing, there are approximately 30 million developers in the world, and this number is growing by about 5% per year.
The total addressable market for traditional code hosting companies alone is in the range of a few billion dollars.
The serverless hosting market is valued at around $10 billion globally and has a very high projected growth rate of between 20 to 25% year-on-year.
Additionally, our product is adjacent to AI but doesn't compete in the space directly, so we expect improvements to AI to further enhance our product.
Currently, GitLib is built by two of us, so myself, Natalie, I'm the apper, and I'm also a software engineer with a background in the security industry, and then Zoran, the shipper, who's a very experienced technologist with a tendency to dig deep into the stack.
We're weeks away from our launch date, so if what we're building resonates with you, either as a potential customer or as an investor, please reach out because we'd love to hear from you.
For updates, please follow me at Natalie Marleny on social media or subscribe to the GitLib mailing list.
Hello, everyone. I'm excited to introduce Hanabi.REST, a product that uses AI to dramatically speed up the development of REST APIs.
Many of you, whether you're developers or executives, have probably encountered situations where you needed a web API for a prototype.
However, setting up a database or configuring the environment often becomes a bottleneck, preventing you from focusing on your core ideas.
That's where Hanabi .REST comes in. By typing just a few lines or spending about a minute providing a prompt in your browser, our platform automatically generates a REST API tailored to your needs.
Now I'd like to give you a quick demo to show how simple it really is.
First, I go to Hanabi.REST in my browser.
Then I enter a short prompt. For example, let's type in Twitter -like API.
When I click the button, the AI spends a moment thinking. After that, it begins generating documentation and code.
Since the output is streamed in real time, there's no dull waiting period.
Now the TypeScript and SQL code has been successfully generated.
The build process is fast and runs in the browser, so I can immediately test the newly created endpoints.
There's also a built-in environment to send requests and inspect the database values.
If I want to deploy this API, all it takes is connecting my Cloudflare account and clicking a button.
From there, the API is automatically deployed to Cloudflare Workers, and it's up and running in no time.
If you're unsure how to use the generated API, simply check the API guide.
In addition to that, we automatically generate extensive documentation, making it easy to understand the API's specifications.
As you can see, the real power of Hanabi.REST lies in its ability to turn your ideas into fully functional APIs with just a few lines of text and a single click.
Because we're a team passionate about Cloudflare, Hanabi.REST integrates seamlessly with Cloudflare's products, offering one-click deployments to Cloudflare Workers and D1 databases.
Moreover, our platform provides a live sandbox right in your browser, generates code optimized for edge runtimes, and allows you to share or fork the generated code.
We even offer a CLI that can be integrated into your existing development workflow, ensuring you have the tools you need to bring your ideas to life.
Best of all, everything is done through your browser, so whether you're on your office computer or a smartphone, you can spin up a REST API anytime inspiration strikes.
If this sparks your curiosity, I encourage you to visit hanabi.rest and hit that Generate button before you know it.
Your ideas will be up and running. Thank you for listening, and we hope Hanabi.REST helps you bring your vision to life faster than ever before.
Hello, hello. I'm Neha, the co-founder of NotHotDog. We are helping make AI agents enterprise-ready.
We are backed by Berkeley SkyDeck, and you can scan that QR to get to our website.
Imagine this. Your customer care agent ends up costing you $260 million because it could not tell the difference between a McNuggets and a milkshake.
That's exactly what happens to McDonald's. And they are not alone.
A lot of enterprises end up spending millions of dollars because their AI agents fail spectacularly in production.
The reason is simple. AI agents are not tested well enough to be ready for enterprises.
Agent testing is manual, fragmented, and extremely time -consuming.
90% of the testing is manual, and there are humans sitting and mimicking other human interactions with systems.
Limited evaluation methods.
There is no systematic way to evaluate agent behavior at scale today.
And there are unknown compliance risks. Edge cases that are discovered only after the agent is deployed to production.
And there is no way to assess these risks before deployment today.
We are solving for that. We are building an agent-first testing platform.
It helps you get 3x faster releases because it's a quick one-time setup.
You sync your agent with our platform, and we automate testing with every update.
Two, we help you achieve 80% higher reliability.
Our co-product helps build 20 diverse personas and generate 100 -plus scenarios for the agent and all these scenarios are run in parallel, helping achieve 80% higher reliability.
In addition, we allow users to also create custom scenarios and bespoke personas using natural language.
We have a wide range of metrics and a dashboard that gives you instant access to any issues in production, track real-time performance of all your tests, and also set up metrics like response time, accuracy, failure rates, in addition to other standard metrics like F1 score, blue score, precision, accuracy, and recall.
We sell typically into the engineering and product teams, but certain cases we're also selling to domain experts.
We've spoken to 135 customers, and every single one of them agrees that manually testing AI agents is extremely painful.
According to Gartner, there would be 115,000-plus engineering teams building AI agents by 2030.
We intend to get just to 16,000 of those teams by year five at a price of $550 per team per month.
That gets us to $105 million in annual recurring revenue by year five.
This is the engineering team that's building this product. I'm Neha, the co-founder and CEO.
I've worked with Zoho for seven-plus years. I helped build and ship their accounting product for the Middle East region and scaled to 100,000 -plus users.
I'm also a Carnegie Mellon alum and part of Inception Studios' founder circle.
My co-founder and the CTO, Akash, is a software engineer at Apple, worked on Apple Intelligence, among certain other undisclosed projects.
He's previously been the founding engineer at a PropTech startup and is also part of OnDeck founder circle.
Our ask is fairly straightforward. We are raising our first round, and we appreciate introductions to angels who are active in the AI agent space.
We want introductions to customers building AI agents because we are now live with our beta product.
We're also seeking advisors, so intros to advisors who are deeply technical and active in the AI space would be very helpful.
If you come across any of these folks, please reach me at my number or email and make that introduction.
Thank you so much. Hello, everyone. I'm Shishir Bashel, the co -founder and CEO of Proma.ai.
Most organizations suffer from shitty workflows that cause constant delays, repetitive mistakes, information silos, and a long list of other issues.
In fact, a research from IDC estimates that U.S. corporations lose more than $5 trillion annually to WordPress inefficiencies.
And that's where we're coming.
Proma is a process management and automation platform that helps you build and run intelligent workflows.
With Proma, you can go from a problem like, how do I manage my team's leave requests and approvals to an intelligent solution with feature-rich forms, actionable Kanban boards, customized calendar views, and aesthetic automations with LLM and other machine learning models, as well as insightful data dashboards, all in less than an hour.
Dr.
Bipul Luthil and myself bootstrapped a B2B SaaS product to 1 million ARR before co-founding Proma.
We both have a strong background in machine learning and AI, and we have now joined forces to solve this $5 trillion a year WordPress inefficiency problem.
We currently have nine beta clients in the various different industries, and we are looking to raise $3 million in seed investment in the next few months.
We'd love to share more with interested clients and investors. Thank you.
Hi, I'm John from WebStudio, which is an open source advanced visual builder. It's an extremely powerful tool, but that comes at a cost.
It takes a long time to learn.
We want to solve this problem with AI. Let's look at the current tools out there right now.
Here's WebStudio along with its competitors. It's at the top because it's a production grade tool.
It's towards the right because it has a high time to learn.
It's not all the way to the right like custom code, but we can do a lot better.
Now, I want to point out these tools down here. They're not production grade ready.
However, we can learn from them because they're using AI to kickstart projects.
The problem is you're left with code at the end of the day. This already eliminates a lot of people.
Plus, many people that can code still don't want to work on that.
How do we get WebStudio all the way to the left? The answer is AI. Let's look at where WebStudio is at today.
We are a production grade product. It is dynamic.
You can connect to any headless CMS out there. We only focus on building the visual builder and highly performant hosting, thanks to Cloudflare Workers, and we are open source and privacy focused.
Some of the numbers, we've served just about 100 million requests in the past 30 days, almost 10 terabytes of data.
We have about 60,000 users and over 100,000 projects with a very active community.
Now, this is where we want to go.
We want to be the first no-code AI platform for designers and marketers.
By using diffusion models for creativity and LLMs for the human interaction with them, we can populate our UI using AI.
This is a very important distinction.
AI will populate a user interface and people that don't know how to code can use the user interface to take the project that last mile.
The total addressable market in 2024, core web development services as a whole, 40 to 80 billion, and website builders between 5 and $10 billion.
The team at WebStudio is comprised of world -class engineers, UI experts with over a decade of experience, and top open source contributors and no-code experts.
Our go -to-market strategy is product-led.
We release often, we talk about these things, we build things that people love and get them talking about it, and we're building a movement.
You can see this on our socials and our very active community. For monetization, we started in 2024, made 180K, and we are now exploring enterprise deals.
And our seed raise, this is where we want to integrate these AI workflows to be able to use this UI but have AI populate it with user input.
In conclusion, WebStudio is a production -grade website builder.
It's very advanced, so it has a high time to learn.
By integrating AI into the platform, it'll make it much more efficient for creators.
Thank you for watching. You