Anon: Your Agent's Passport to Every Site on the Internet 🌐

CV Deep Dive

Today, we’re talking with Daniel Mason, Co-Founder and CEO of Anon.

Anon is a developer toolkit for building user-permissioned integrations on the Internet. Founded by Daniel Mason and his co-founder Kai Aichholz in 2023, Anon supports AI companies in creating robust and reliable integrations that require user authentication. The company’s mission, as explained by Daniel, is to enable AI agents to act on behalf of users in a secure and seamless manner, especially in cases where formal APIs are not available.

Today, Anon supports integrations across consumer sites such as LinkedIn, Instagram, Twitter, and enterprise platforms, making it easier for developers to build and scale their AI-driven applications. Notable customers include developers at MultiOn, Personal.ai, Elevate, and more. In April 2024, Anon announced a $6.5 million round closed at the end of last year involving Union Square Ventures and Abstract Ventures, among others. 

In this conversation, Daniel takes us through the founding story of Anon, his vision of building an AI agents platform, and Anon’s roadmap for the next 12 months.

Let’s dive in ⚡️

Read time: 8 mins

Our Chat with Daniel 💬

Daniel - welcome to Cerebral Valley! First off, give us a bit about your background and what led you to found Anon? 

Hey there - I’m Daniel. I’ve been a founder for many years now, starting companies since my early twenties - and for the last 7 years, I’ve been building venture-backed startups. Prior to Anon, I co-founded a business called Spring Labs, which is a data security and identity company that builds PII tokenization, among other products for the financial services sector. We served early-stage fintech companies up to larger institutions like credit bureaus and card networks. TransUnion became a customer and our largest investor, leading our Series B round in 2021. 

I spent four years building that company, and we raised about $68 million and built a team of 40 to 50 people. I left Spring Labs in 2021 and worked in venture capital for a year at a fund called Framework Ventures as a partner doing investments. I then started Anon a little over a year ago in March of 2023. 

The inspiration for Anon came from an AI agent product my co-founder and I built early on, which was also called Anon. We built an AI agent in our users’ Gmail inboxes by reading message context and suggesting actions.. It was a night-and-weekend experiment that blew up a bit, and we onboarded a bunch of users.

Most user requests involved asking us to take actions on their behalf, which required authentication on sites without public APIs. People wanted us to respond to messages in their LinkedIn inboxes, apply to jobs on Indeed, make dinner reservations on OpenTable, plan vacations, synthesize news articles, and more. We found that in 95% of cases, there were no APIs for these actions. We realized that if AI is going to behave as a human online and act as extensions of us for personal, professional, and enterprise use, it needs a way to get authenticated access to the Internet, just like humans do. The web is not designed for agentic behavior in mind. That’s the inspiration behind Anon.

Give us a top level overview of Anon - how would you describe the startup to those who are maybe less familiar with you? 

Anon is a developer toolkit for building user-permissioned integrations on the Internet. We’re building the connective tissue between AI agents and a user’s identity platforms. We’re mainly targeting AI companies because we believe AI represents a new paradigm in delegating identity online. However, we also have customers who aren’t AI companies but are building these user-permissioned integrations on their own. You could compare our business to Plaid in some ways. Plaid built APIs for banking institutions that didn’t have them, giving fintech companies access to that data with user permission. They coined the term “user-permissioned integrations.” We’re doing the same for the AI industry.

In essence, the user logs in and grants the agent access to their account, even when there’s no formal API offered by the service. It uses the user's account credentials to log in, gain access, take actions, or pull down data—whatever the agent is doing on the user’s behalf.

Who are your most user’s today? Who’s finding the most value in what you’re building with Anon? 

We’re working with early-stage or mid-stage companies that are building AI agents, which makes them great customers for Anon. The biggest considerations for us right now are what use cases are you supporting and what integrations do you need? 

We currently support agents built on both web and mobile. We have specific support for Web / Chrome and iOS via Swift and React Native SDKs.  While we’re constantly expanding to other platforms, our initial focus has been on web and iOS.

We also have a list of integrations in our developer docs. We started with many of the most popular consumer sites on the Internet, such as LinkedIn, Instagram, Twitter, Uber, airline sites, OpenTable, and Resy, among others. We’ve noticed that many enterprise users especially want to be able to automate workflows in LinkedIn so this is a key integration we offer. For example being able to automate actions such as sending messages, adding connections, and researching users or pages. We also support enterprise integrations, and by the time this story comes out, we’ll probably have added many more integrations to our docs. We list the most popular integrations in our docs but always encourage potential users to reach out and tell us about their use cases because we continually add integrations based on user requests.

We add new integrations based on customer demand. If you come to us and say you need access to certain sites, whether they’re deep enterprise sites or edge-case consumer sites, we can typically build those integrations very quickly. We have an internal tool for building integrations that we plan to eventually make available to customers, allowing them to use our platform to build their own integrations. For now, we handle integration building based on what our customers need.

The world of AI agents has become a very sought-after space for a lot of new and established players in the AI ecosystem. What would you say differentiates Anon today versus others in the space?

Totally. It’s interesting. We come across hundreds of companies building agents in some form or fashion. One thing that stands out for us is our belief in authentication as a guiding principle for an automated internet. There is a growing field of infrastructure providers in the AI space. We haven’t seen anyone who does exactly what Anon does from an authentication perspective. A lot of companies have tried to build functionality like Anon’s in-house because it’s a core part of some integrations, especially if you’re focusing on niche sectors like automation for SDRs, account executives, or customer support, and you need to use LinkedIn all day on behalf of customers.

We’ve built our integrations in a superior way to building them in-house because non-API integrations are inherently brittle. Since we maintain integrations on behalf of many customers, we are usually able to deliver better reliability with way fewer engineering hours than in-house solutions. We offer strong Service Level Agreements for uptime and reliability. We’ve invested heavily in technology that makes these integrations work seamlessly, especially if you need to build many integrations. Doing it yourself would burn a lot of engineering cycles. 

I think our approach to data security is also a huge differentiator. We’ve built this business so that even deep enterprise companies are satisfied with how we handle PII and user data, which is a big differentiator for us. This is an area that we are continuing to invest in to provide services to larger enterprise customers.

How do you measure the impact that Anon is having on your key customers? Any customer success stories that you’d like to share?  

We’re going to be making a lot of customer announcements in the coming weeks, so I’ll hold on a bunch for now, but one customer we’ve already announced is MultiOn, a consumer agents company. If you’re using MultiOn’s API product or their web playground products, you’re interacting with Anon out of the box. There’s a tab on the services page of their playground where you can see all the integrations we’ve built that they’re currently using. Companies in that sector, which are building various verticalized versions of personal assistant apps or apps with AI taking actions on your behalf, are really where we have a strong product. It’s a really exciting area for us.

I highly recommend checking out their site. It’s cool. You can say things like, “Book a restaurant table for me,” “Book a flight or Uber for me,” or “Send a message on LinkedIn for me,” and you’ll see our SDK in action repeatedly. Anon authenticates each identity platform and enables MultiOn’s agents to get supercharged. Get out there and try their web product to see how it works.

What's been the hardest technical challenge around building Anon thus far?

Our technical challenges are somewhat different from other companies in the AI space because we're not an AI company, we're an infrastructure provider for AI companies. I think the biggest challenge is building robustness and reliability among integrations that are supported via user permissioning versus supported via formal API. The opportunity we are building for is so large; that is, we’re building for automations for apps that currently don’t have formal APIs but that everyone would want to have linkages into.

So I think this is the reason it's hard to build internally is because these integrations can lack reliability. You have to build a huge amount of infrastructure around IP, proxying, around how you handle two factor authentication. And of course we think a lot about how you make the experience feel very seamless for end users and handle communication with end users around their core workflows. There are  a lot of considerations at the intersection of user experience and reliability that present. I think that's probably the hardest part of building our company and we expect to be the hardest part moving forward. But that's the reason it's valuable for developers to work with instead of trying to build this on their own.

How do you plan on Anon progressing over the next 6-12 months? Anything specific on your roadmap that new or existing customers should be excited for? 

Today, we’re in a closed beta onboarding individual customers. We have a really large waitlist and we’re onboarding a few each week to make sure the product is working well and to gather feedback efficiently. In the next few weeks, the plan is to move to open beta, where you can sign up on our site, start using the integrations, generate API keys, and everything will be handled automatically from the user onboarding and signup process. The big goal we’re working towardsis to get to open beta, onboard everyone from our waitlist, and go into hyper growth in user acquisition and developer usage.

One of the things this will enable us to do is invest in the developer ecosystem. For example we’ll be able togo to Cerebral Valley hackathons and offer prizes to developers who use our product as part of their project. We really want to reach grassroots developers, people building side projects, and support big venture-backed startups as well as everyday developers. We think there’s a lot of magic that comes from making our product available to these kinds of users. [If you’re a developer who’s interested in Anon, DM me on Twitter and will personally help you onboard!]

Looking further ahead, we’re also focused on supporting all platforms, no matter what browser or device you’re using to build your app. We want to expand beyond the web and iOS SDKs that we currently offer. We also want to increase our catalog of integrations significantly, from the 12 to 15 we have now to hundreds. Our goal is to create a “build your own integration” tool, where developers can build something unique and open-source it.

Our dream is a future where you can add new integrations to our catalog, and build our a developer marketplace around actions that you build and contribute. We want this to be a community-driven project, leaning into the open-source aspect. We want to be the go-to catalog for AI companies looking to integrate with any site on the Internet, and that’s how we plan to grow from where we are today to where we want to be in the future.

Lastly, tell us a little bit about the team and culture at Anon. How big is the company now, and what do you look for in prospective team members that are joining?

We’re eleven full-time employees today. We're based in San Francisco, with an office in Hayes Valley. We work in person every day, which is a big part of our team culture. Interestingly, the team was remote for the first six to eight months. After we raised our seed round last fall, we decided to move everyone to San Francisco and build the company in person. It’s paid off big time in terms of productivity, the enjoyment of building the company, and access to customers and the ecosystem. It’s great to be able to walk down the street to a hackathon or a customer’s office and integrate in person.

The team is mostly engineers—seven of the eleven are engineers. We have one person in product, one in design, and then myself and one other person handling the business aspects. Everyone on the team has written code at some point. Since we’re building a developer product, a strong technical background is crucial for us. As for who we look for, we love hiring former founders. Almost all of our early hires were former founders or CTOs, people who have been through Y Combinator and held leadership roles. We want people who are big on ownership, self-motivated, and really passionate about building and owning their work without needing a lot of guidance.

We also like to have fun while we’re doing this, which is why being in person is important to us. We do a lot of team lunches and events, like a team hike over the weekend and playing board games. We’re a close-knit team and value having people who want to lean into that kind of culture. We’ll be hiring soon; we just added four people in the last four weeks, so we’re taking a short break on hiring. But for anyone who’s an incredible backend or full stack engineer, we’re always open to having those conversations.

Anything else that you’d like readers to know about you or Anon? 

The last thing I’ll say is that if you’re building in the AI space, feel free to shoot me a message! I want to personally help you build valuable integrations & agent automations. Tell us what you’re missing. This space is evolving so quickly, and it feels like every time you blink, there are five new announcements, and things have changed. We’re just trying to keep pace with it all, and we do that by connecting with leaders in the space. We want to play a big role in AI enablement. We’re not an AI company and we’re not your competitor. If you’re building something in AI, we’re here to help make your product better, make your users happier, and extend your functionality.

We’re excited to work with developers in any way we can to help them build the best AI apps on Earth.

To get started with Anon, visit docs.anon.com and click “Get Access”.  If you have any questions about our product or want to chat, send a DM to Daniel on Twitter or email us at [email protected]. Thanks and excited to see what you build using Anon!

Conclusion

To stay up to date on the latest with Anon, follow them on X and learn more about them at Anon.

Read our past few Deep Dives below:

• Zilliz Cloud is powering vector search for the enterprise 🔋

• Kolena aims to redefine quality standards in ML 🎛

• Pinecone Serverless is turbo-charging your time-to-prod ⚡️

• Cleanlab is automating data curation at scale 🌐

If you would like us to ‘Deep Dive’ a founder, team or product launch, please reply to this email ([email protected]) or DM us on Twitter or LinkedIn.

Join Slack | All Events | Jobs