Since May, Survol has been built more or less in public on this blog: the principles, the screens, the trade-offs, and a few things we gave up on. One thing was missing, and it is the only one that really counts - teams using it for real. Hence this announcement: the enterprise beta is open, to ten organisations.
The terms fit in two lines. €5 per user per month instead of €29, and a direct line to the founder - not a support channel, the person who writes the code. Ten seats in this first wave, an answer within five working days.
Let us say straight away what any announcement of this kind should contain and almost never does: as of today Survol has no paying customer and has never been proven at a third party. Everything you have read here describes a product built with conviction and tested by one person. That is exactly what this beta is meant to settle.
Five euros, and why not zero
The question is a fair one: at this stage a free beta would be easier to fill. That is precisely the problem.
A few euros a month fund nothing. They filter.
A free beta attracts curious people who open the tool once and never come back. A symbolic amount is only paid by teams seriously considering adopting it - and those are the only ones whose feedback is worth anything. It is the most honest filter we have, and it costs less than a coffee a month per person.
One clarification that matters when a price is involved: those €5 are Survol's price, not your total cost. We do not resell tokens. Your coding-agent plan stays yours, with your provider, at the price they charge. If you prefer the managed mode, where Survol runs the agent on its own servers, tokens are billed back to you at the provider's public price, at no margin.
Organisations in this first wave keep preferential terms at commercial launch. The beta price is not a launch discount that closes after three months.
What "a direct line to the founder" actually means
Just about everyone uses that phrase, so here is what it covers here: your feedback reaches the person who writes the code, with no ticket, no queue and no intermediary rephrasing it. A bug reported in the morning can be fixed in the afternoon. A substantive objection opens a discussion, not a canned reply.
The trade-off is arithmetic, and it is the real reason behind the number ten: beyond that, it stops being a direct line and becomes a queue that does not admit to being one. We would rather have ten organisations genuinely supported than fifty waiting for an answer.
In the same spirit, the next pieces of work are decided with the pilot teams. Survol's roadmap is currently one person's roadmap: it needs to be argued with.
Who it is for - and who it is not for
The profile is a narrow one: a team of 5 to 100 people already using a coding agent, which is finding it has lost track of what that agent produces. What is done? In which version? On which branch? Who validated what? If those questions sound familiar, you are in the right place.
Two cases where you should sit this one out. If you are looking to try a coding agent, this is not it: Survol does not replace your agent, it sits above it. With no agent, it has nothing to steer. And if you have no real product to hand it, the beta will teach you nothing - and will teach us nothing either.
What we ask in return
- A real product to steer - new or existing, it does not matter.
- An active coding-agent plan, or the wish to try the managed mode.
- Regular feedback: thirty minutes a month is enough.
And one thing that is not on the form: patience. You will run into bugs, unfinished screens and at least one decision you find questionable. That is what a beta is. If it were not, we would already be selling the product at €29.
What we are not promising
Three limits worth knowing before you apply rather than after.
The scope of this first version is deliberately narrow: one coding agent and one source-control provider. The architecture is designed to take others - the code names no provider - but until the loop is proven end to end on a first pairing, adding a second would double the surface to validate while proving nothing more. If your code lives elsewhere, this is not the right moment.
The savings figures you may have read here are hypotheses, not measurements. Survol tracks token consumption per version, so it will be able to compare execution strategies on real features and show a measured saving, customer by customer. That is what this beta has to produce. Not one more promise.
The roadmap can move, and some of what has been mocked up will not survive contact with a real team. We would rather say so now.
Applying
The form is eight fields long and lives on the home page: apply to the beta. Ten seats in this first wave, an answer within five working days. Your answers are used solely to select the pilot teams.
In short: ten organisations, €5 per user per month instead of €29, a direct line to the founder and a say in the roadmap - in exchange for a real product and thirty minutes of feedback a month.
If you are hesitating because your situation fits none of the boxes above, write anyway: the "your context, in a few lines" field is the one we read first.
