Flamingo Looks To Build Autonomous MSPs With Open Source, AI And $4.5M In New Funding

Flamingo CEO Michael Assraf wants MSPs to trade security tool sprawl and ticket queues for AI agents, open-source infrastructure and automation and is backing that mission with a new $4.5 million seed round to push the startup toward its imminent commercial launch.

Flamingo, a startup building what it calls an open, AI-native infrastructure layer for IT and security, is moving toward commercial launch with fresh seed funding and a platform aimed squarely at MSPs. The Miami-based company this week is announcing a $4.5 million seed round led by Vertex Ventures, bringing total funding to $6.7 million.

Flamingo’s OpenFrame platform combines an open-source-based tooling stack with AI agents designed to automate IT service delivery, security operations and end-user support across a unified API and data layer. OpenFrame Core is slated to consolidate 19 IT and security product categories, with the first generation already covering such areas as RMM, MDM, monitoring, automation, remote access, patch management, security monitoring and ticketing.

CEO Michael Assraf told CRN Flamingo is a single platform that brings open-source IT and security tools together for MSPs and adds AI agents to operate across them.

[Related: OpenMSP And Startup Flamingo To Offer MSPs Open-Source Alternatives To Expensive Tools]

“Flamingo is a combined product for a complete IT security platform based on open--source tools as one unified solution for IT teams with AI agents on top of it to orchestrate all those tools in a very transparent and unified manner for ITSM [IT service management], service delivery and so on,” Assraf said.

Those agents are central to Flamingo’s MSP goal to reduce routine ticket handling while still preserving approval workflows for more sensitive actions, Assraf said.

“Our technician-side AI agent is also capable of running more privileged operations and operations that go across the entire fleet,” he said. “So it’s not segregated only for a specific device. Mingo, which is the technician-side agent, is also able to hunt for problems proactively and fix them and auto-suggest the user to fix them rather than waiting for a user to complain.”

Assraf said Flamingo grew out of his earlier experience with MSPs at Vicarious and his view that open source and AI together could change MSP economics.

The new funding is meant to help Flamingo move beyond beta, stabilize the product, and accelerate AI development as it prepares to start billing customers. Assraf said the company onboarded about 400 MSPs in beta and wants to launch only when the product delivers a much more modern MSP and end-user experience.

“I felt like we needed more R&D power in order to accelerate the implementation of our vision,” he said.

There’s a lot going on at Flamingo. To learn more, read CRN’s conversation with Assraf.

What is Flamingo?

Flamingo is a combined product for a complete IT security platform based on open-source tools as one unified solution for IT teams with AI agents on top of it to orchestrate all those tools in a very transparent and unified manner for ITSM, service delivery and so on. The product is made up of 19 different tools under the hood for every category you can think of in terms of IT security. We have RMM, PSA, our own SIEM [security information and event management], a lot of different components. We are currently in Gen 1 where we have eight modules, and by the end of the year we’re going to have all 19 available. A lot of those tools are advanced, like zero-trust and identity and access management and things like that. But the core tools are already out.

On top of it all are two AI agents. One AI agent sits on the end-user computer and acts as an IT helper. So if the user has any issue, it can communicate with our IT desktop app. If it’s something easy to fix and doesn’t require technician approval, the IT helper, her name is Fae, will automatically fix it for the user without even opening a ticket and all this big mess. If more privileged access is required to do more extensive stuff on the device, Fae will ask for approval from our technician-side agent called Mingo, and Mingo will ask for approval from the actual user. Our technician-side AI agent is also capable of running more privileged operations and operations that go across the entire fleet. So it’s not segregated only for a specific device. Mingo ... is also able to hunt for problems proactively and fix them and auto-suggest the user to fix them rather than waiting for a user to complain.

Our goal is to create a completely autonomous MSP with anomaly detection data streams, user reporting problems, and everything getting fixed completely automatically. The user can configure alerts and guardrails and will get alert and guardrail approval requests to their mobile phone. We have a mobile app, and we have a desktop app that was recently released. The mobile app is still pending approval by Apple and Google Play, but it’s ready. So that’s how we see the MSP and their IT team’s problem of continuous delivery and answering trouble tickets when needed.

Can you give an example?

Imagine a user needs to reset their password. To reset their password, the MSP configured the guardrail to require technician-side approval. The user will report, ‘Hey, I’m logged out of my account. I need to reset my password.’ A notification will be sent to the MSP technician’s phone. The technician will hit ‘approve,’ and the password is going to get reset, and the ticket will be automatically closed. You’re going to have a record on our dashboard, but it’s completely transparent in terms of all the hassle that we used to have with opening a ticket, closing a ticket, handling work time, all this type of stuff.

Flamingo’s technology is aimed specifically at MSPs?

Yes.

What are MSPs missing today that caused you and your team at Flamingo to develop this technology?

That’s a good question. Flamingo is my second startup. I started a cybersecurity company Vicarious back in 2016. It was a vulnerability remediation company. And most of our success in the U.S. came from MSPs. This is how I learned about this industry. This was before the AI revolution happened. I saw a lot of younger, smaller MSPs had enough with paying vendors and reselling tools, so they started switching off to open source and serving it as white label to their users. Most of the time, their users were brand-agnostic. So I started wondering if people could build a very highly efficient and profitable MSP just by relying on open-source tooling rather than paying high vendor cost. That could cut around 30 percent of the expenses that MSPs have, bringing the MSP from, I don’t know, 20 [percent] to 30 percent gross margin to 50 [percent] to 60 percent, which is a very interesting business. And then AI came. If you open a common [profit and loss report] for an MSP, you’re going to have 30 percent profit from software reselling. They will buy RMM and then deploy it for the users. You also have very high costs for labor. I don’t necessarily think labor should be decreased or removed, but you can serve many more customers with the same people with AI.

So the combination of open source and AI gave me the idea and the vision for Flamingo. I started by researching very deeply whether something like this was even feasible. Are there enough open-source tools in the market to build this very broad open IT security operating system? Can you build on top of it all the AI features that will serve in a good manner the MSP users? One of the fruits of this research is our community called OpenMSP. OpenMSP is a vendor directory that lists commercial tools and open-source tools and how you can switch from commercial tools to open-source tools. Anyone can join. We have almost 700 MSPs on our Slack community, and we have around 1,000 registered to OpenMSP. So it’s a very interesting stream that we identified in the MSP space with this open source. But to be honest, it’s not something a lot of bigger MSPs do. I found it very common for smaller or younger MSPs.

You had some news. What’s going on?

We had our pre-seed round happening in February of last year, and with the launch of OpenMSP we also started to collect waitlist signups for our product. We got to around 3,000 validated MSPs on the waitlist, and my sense was that we were getting to the point where we have product-market fit, and we need to raise more capital in order to make the product more stable and to add more AI features. Our product is self-serve. Also, our pricing is completely open. Anyone can go on the website and see the per-device pricing and also the AI consumption-based pricing. I felt like we needed more R&D power in order to accelerate the implementation of our vision. So we went out for a very intense funding process of around three weeks, and we got a very good lead with Vertex Ventures. ... We completed the funding round for $4.5 million with the participation of our existing investors. This allows us to really kick off our commercial operations. The product is still in beta. We’re launching this month. We want to go out when the product already offers 10X the experience for MSPs than what they have today. So things like real-time notifications from AI, solving things directly with AI, desktop applications that MSPs can use, a 2026 experience for end users, and a mobile app for end users. All of these are things we want to complete before we go commercial. And from what we’re hearing from MSPs, we have 400 MSPs on-boarded to the product in beta. We think that most of them will move with us once we start billing. We already launched the billing and usage section on the product, so they can see how much they’re going to spend before we kick off the billing. …

We have our Gen 1 ready to go. We’ll come out with Gen 2 and Gen 3. Gen 3 will include all 19 different modules. We’ll also be adding more and more stuff on the AI, on the mobile app, on everything that will make the experience of running an MSP and consuming MSP services a 2026 AI-native experience.

Which modules are all ready to go with Gen 1?

Right now we have RMM, remote access control, patching, PSA, SIEM, MDM and documentation. We have remote access from the SOC [Security Operations Center] and the user perspective. They are ready to go.

What comes next?

Next will be cloud backup, EDR, EPP [endpoint protection platform], identity and access management, network monitoring and secure email gateway.

Where does Flamingo do its development?

All over the world. That’s our end goal. We want all the tools to be speaking in a unified manner. In terms of R&D, we have about 35 employees in 12 countries. We are all remote. ... We are currently in the process of acquiring SOC 2 Type 2.

What happened to your previous company, Vicarious?

It’s still running and doing well. I liked what I saw going on with AI, and I said to myself I have to start something in this space because it’s just massive. It’s like how I saw the iPhone revolution. I think that AI is 10 times bigger. I think AI is as big as the industrial revolution. That’s how it’s going to change the world. So I had to do something in this space, and I always liked the MSP industry and IT. It’s something I’ve always been part of throughout my career. So I decided to move on.

What is the base on which Flamingo’s AI technology is built?

Our AI is a layer on top of existing frontier models. You can choose which model you want to work with. We recommend Claude. We used to support Gemini. Now we support OpenAI and Claude, but we are model-agnostic. We sit on top of the AI model.

Why recommend Claude?

With all of our benchmarks, with all of our testing, Claude is the best model for operating in the IT security space. For our purposes, it performs much better than any other model by far. It’s not even close.