The first customer arrived, and the software fell over. Sandro Munda had built Forest Admin’s earliest version for a single user, then occupied the only available place himself. During that first onboarding, he had to restart production in front of the person he was trying to impress. The customer ended up with database ID 1. The founder got ID 2. A small humiliation, neatly indexed.
It is a pleasing origin detail for someone whose working life has revolved around the parts of software customers rarely applaud. Admin panels. Permissions. Data connections. The machinery behind the polished screen. Even the founder’s embarrassing entrance happened backstage, in the database, where the evidence could be stored indefinitely.
Munda went on to lead Forest Admin for a decade. In July 2025, he announced that Maxence Bruyas would take over as CEO, while he remained a founder and board member. He described a business generating millions in annual recurring revenue and serving hundreds of companies. Then he went back to building. His next company, RootCX, concerns another problem that becomes visible once the demonstration is over.
The question that chose the framework
Before Forest, Munda was a freelance developer making custom applications. A recurring question shaped his choices: would the client need an admin panel? Sometimes the answer sent him toward Ruby on Rails because RailsAdmin supplied a convenient starting point. The customer’s internal interface was helping decide the technology for the whole application.
Consider the oddity. A business commissions software to do something particular, yet a familiar maintenance requirement can end up steering how it is built. Every organisation has its own processes; each still needs people to inspect records, make changes and carry out everyday work. The repetition was there, even when the business details differed.
Munda’s early Forest writing challenged the assumption that those differences made a shared product impossible. He wanted more care given to the interfaces employees used behind the scenes. Consumer applications had trained people to expect considerate design. The software waiting for them at work was often allowed a rather more austere personality.
He developed Forest with eFounders, the startup studio now called Hexa, and co-founder Arnaud Besnier. The project was underway by 2016, when Munda was publishing his argument for offering admin panels as a service. A freelance irritation had become a product proposition: let developers adapt common machinery to their business, while spending less time constructing it.
A company behind the curtain
The appeal was concrete. Forest supplied an administration interface for managing application data and operations, without requiring a team to build the entire front end itself. In a 2020 interview, Munda compared the product to Lego bricks for developers to assemble. The comparison made room for both repetition and invention. Standard pieces could still produce a specific result.
The early company attracted Connect Ventures and backers including Xavier Niel and Taavet Hinrikus. Its seed financing was announced in early 2018. In November 2019, Forest announced a $7 million Series A led by Notion Capital and Runa Capital. Money was following an idea whose visible output might be a workaday administration screen. Venture capital occasionally wears sensible shoes.
At that stage, the company was planning to strengthen its relationship with developers, broaden integrations and make the platform easier for business teams to use. Those priorities reveal the balancing act in Munda’s proposition. Developers needed a tool that fitted their applications. The people operating those applications needed something they could actually work with.
There is a temptation to tell a software company’s story entirely through its funding rounds. Here, the more revealing object is the back office itself. It sits between the people writing the code and the people carrying out the business. Munda had chosen a product that had to earn the confidence of both.
The lumberjacks had paperwork
Munda also wrote about the organisation around the code. In 2018, he published a recruiting essay titled “Hiring is a two-way street.” Candidates, he observed, were expected to explain their accomplishments and suitability. Companies ought to make a corresponding effort to explain why anyone should join them.
Forest’s internal language had a woodland theme: employees were lumberjacks, with a published Code of Honor. The name supplied the joke; the accompanying argument was quite serious. Publishing values made the company accountable for preserving them as it grew. A culture statement was something employees could later hold against its authors.
His recruiting pitch also acknowledged a difficulty hidden inside an enormous market. A product for many kinds of web businesses could spread its makers too thin. He advocated methodical targeting and documented processes. For a founder proposing a broadly reusable tool, the discipline of choosing where to concentrate mattered as much as the breadth of the ambition.
That combination gives his public writing its particular texture. He can be expansive about the opportunity and fussy about execution within the same argument. The administrative instinct reaches beyond the product. If the company is going to help other people organise their operations, it should probably know how its own operations work.
Code with a past
By 2021, Forest had roughly 30 people, including 20 developers led by Besnier. Its product had to work inside customer environments that its own engineers could not always inspect directly. Building something reusable meant encountering the accumulated decisions of many other software teams. Compatibility was a daily job.
Munda’s lesson from that work was to resist wholesale rewrites. Complicated code could embody hard-won knowledge of unusual setups. Replacing it meant risking the loss of solutions whose purpose only became obvious when something broke. The instinct is almost archival: before clearing the room, find out why the furniture is there.
His wider developer record supplies another view of the same person. On GitHub, he maintains a public identity as SeyZ. His projects include jsonapi-serializer, a Node.js library for converting data to and from the JSON API format, and Baboon, a tool for detecting source-code merge conflicts. There is a consistent interest in making awkward technical work less awkward.
His Stack Overflow profile contains years of questions and answers about Git, Python, Node.js and other everyday tools. Alongside the founder biography sits the working programmer’s trail: practical questions, particular failures, useful answers. A company title can compress a career into a line. A developer’s public history tends to leave the screws on the table.
Leaving the CEO chair occupied
The July 2025 handover had a different sort of precision. Munda named his successor and specified his own continuing place as founder and board member. Forest would keep operating under Bruyas. Munda would pursue another venture. The announcement separated the company’s next chapter from the founder’s next job.
His assessment of Forest was understandably personal. He characterised its beginnings as an attempt to create a category before there was an established market around it. By the time he handed over, he could point to paying businesses and substantial recurring revenue. Those were his reported measures of progress, rather than a public financial statement.
He had also been reconsidering what building a software company required. In a separate July post, he argued that easier product development increased the importance of distribution, narrative and a repeatable route to customers. The programmer who once noticed a repeated coding task was now describing a world in which writing code had become easier to repeat.
“The craft moved up the stack.”Sandro Munda, July 2025
Another app, another island
RootCX brings him back to the internal workings of organisations. Its stated purpose is to provide shared infrastructure for internal applications and AI agents: the database, identity controls, permissions, deployment and records of activity needed when software enters daily use. These are product commitments, with the practical details carrying much of the argument.
In April 2026, Munda described AI-generated applications as islands. A team might produce a CRM, a billing dashboard and a task manager quickly, then inherit separate places to store data, separate logins and separate systems to maintain. The speed of creating an application could multiply the work of looking after it.
The concern extends an older preference. Forest reduced repeated construction around admin interfaces. RootCX proposes a common foundation beneath multiple tools. The technologies have changed, but the recurring expense still interests him: useful work surrounded by infrastructure that somebody has to build again.
He has argued for combining the integrity associated with enterprise systems and the adaptability of modular software. In his account, customer context should survive movement between sales, support and finance. It is an ambition for coordination, and it places a demanding obligation on the underlying software: all those pieces must agree about what is happening.
RootCX’s published principles also put code ownership and self-hosting in view. Its source-available licence converts to Apache 2.0 after two years. Readable code, the ability to operate the system yourself and shared security controls are part of the offer. For a founder selling the convenience of common infrastructure, the terms of dependence deserve attention too.
There is a continuity here worth following beyond any prediction about AI. Munda keeps returning to the work that begins after someone says the product is finished. Who can use it? Where does the information live? How will the next application connect? The first Forest customer got ID 1. Years later, the founder who took ID 2 is still thinking about what happens when another user arrives.