First Version
£0 build feeWe select a few ideas with one clear user and one useful workflow to show what LumenFlow can do. We agree what it will do and when you will get it first. There is no build fee; you approve any hosting or paid API costs.
Bring an idea, a job your team still does by hand, or software that's stuck. We shape, build, test and launch the first useful release. LumenFlow is the system we built to put AI agents to work quickly, with each tool call it governs checked against the job's limits. You make the key decisions. We do the work and hand over the custom code.
A few sentences are enough. We turn them into a clear first version, agree what success looks like, then handle the design, code, testing and launch. You make the important decisions; we do the work.
We select a few ideas with one clear user and one useful workflow to show what LumenFlow can do. We agree what it will do and when you will get it first. There is no build fee; you approve any hosting or paid API costs.
For a larger idea, we map the key user flows, technical approach, first release and build estimate. Take the plan to us or anyone else.
We build, test and launch the agreed release. You approve it against clear acceptance criteria and receive the custom source code.
Start with a free Rescue Check. We read the code and configuration, never your live data. Within 3 working days of getting code access, you get a written answer: keep, harden or rebuild, with the next steps in order.
How the Rescue Check works →A written report: the verdict with its reasons, the biggest risks in order, and what fixing them would take.
We fix the agreed problems and get the software ready to use, at a price and date agreed from the report before we start.
The booking screen above is a browser sample. Separately, imagine we assigned an AI worker to build that feature. Pick a tool request it might make. The real LumenFlow kernel checks a sample policy live and returns what the job would allow, pause for approval or deny.
This evaluates three fixed sample requests. It does not generate code, run a tool or demonstrate the separate Linux sandbox.
You'll see the kernel's decision before any tool could run.
This live check only evaluates permission. It runs no tools, changes no code and publishes no release.
We use LumenFlow while building LumenFlow and LumenFlow Cloud. Its kernel checks each governed tool request before the tool runs. We bring that system to your project and run it for you.
See LumenFlow →Yes. You do not need a spec, prompt, prototype or existing app. Tell us who it could help and what you hope it will do. We can work out the first useful version together.
The first conversation and the Rescue Check are free. We also select a few small ideas for a First Version with no build fee for one agreed workflow. You cover any agreed third-party costs, such as hosting or paid APIs. We share your project publicly only with your separate permission. Other work is priced in writing before it starts.
No. LumenFlow is our delivery system. We set it up, run it and manage the work. You tell us what you want, make the key decisions and receive the finished software.
LumenFlow checks a governed tool request against the job's rules before that tool runs. On supported Linux machines, subprocess tools run inside an operating-system sandbox; other built-in operations enforce the authority LumenFlow gives them. The evidence store attempts to record requests and results and reports trace write failures. We agree your project's controls and review its release against the job you approved.
We aim for the smallest useful release first and show you working software as it takes shape. We give you a delivery date for your build in writing before work starts. A Rescue Check arrives within 3 working days of code access.
You receive the custom source code and access we create for your project, whether the agreed build is free or paid. The written agreement spells out those rights and identifies any third-party or pre-existing components and their licences before work begins.
hellm.ai is Hellmai Ltd, a UK software company. We make LumenFlow and LumenFlow Cloud, and use them in our own work.
A few lines are enough. Tell us the idea or what your existing software needs to do. A person reads it and replies.
Prefer your own email? Write to hello@hellm.ai.