—Surface
Whatever your crew shot, from LAS file to a report your operator can question.
production logs, GLV surveys, radial cement bond logs, injectivity surveys and dip-in pressure surveys.
Depth control, spinner calibration, valve status, zone rates and the write-up — one pass through one piece of software, instead of four tools and a day of assembly.
- Rates, not shares
- Zones come out in Mscf/D, because the spinner is calibrated instead of normalised
- Injection depth, settled
- Which valve is actually taking gas, whether the well is multipointing, and where the tubing is leaking
- A report they can question
- Live charts under your name, and an assistant that answers from the recorded reason
A survey that cannot disagree with the allocation is not worth running.
—The survey
A finished gas lift survey from the demo well
An unresolved valve is a run you can plan. A valve called wrong is a workover you did not need.
—Position
Most interpretation software produces a number for every cell.
This one says so when the data will not carry an answer, and tells you what the next run would need to settle it. Speed is not the pitch. The pitch is that you can see which numbers on the report are measured, which are inferred, and which the well would not give up.
A refusal is a finding
An unresolved valve tells you what the next run has to carry. A guess looks identical to a measurement once it is printed, and costs you the well.
The arithmetic is checkable
Every rate traces back to a curve and a calibration you can open and look at. Nothing on the page is a number the software will not show its work for.
Confidence, per interval
Where the density is measured you are told so. Where it is inferred from the gradient you are told that too, and how much that costs you.
—The job
What it does with your files, in the order it happens.
- 01
The files land
Any LAS, whoever shot it
Down passes, up passes, stationary stops, surface recorders, gauge exports. Curves are recognised on arrival whatever the tool called them and whatever order they came in, so nobody spends the morning renaming columns to match a template.
- 02
The log goes on depth
Collar by collar, and you approve it
Passes are aligned on the casing collars, with shift and stretch recorded separately so anybody reading the job can see what moved and by how much. Your analyst approves the alignment before a single rate is computed from it.
- 03
The spinner gets calibrated
The step almost nobody does
Most interpretation normalises: fastest zone is 100%, everything else a share of it. FloTrace fits the calibration passes your crew already shoots — rps against cable speed — and solves the tool constant and the bearing-friction threshold from them. That is the difference between a percentage and a rate.
- 04
Valves get called
Open, closed, or unresolved
Every mandrel kicks the spinner whether or not the valve is passing gas. What decides it is whether the trend comes back afterwards or settles somewhere new. Where the flowmeter sat inside its own deadband, the valve is reported unresolved instead of guessed — so you know what the next run has to carry.
- 05
Rates come out in Mscf/D
Not “this one is 100% and that one is 40%”
Each zone converts at its own pressure and temperature, so a deep valve yields far more surface gas per downhole cubic foot than a shallow one. Where a capacitance curve exists, the phases are split rather than lumped together as liquid.
- 06
The report comes out
Including the write-up
Log presentation, gradient plots, valve summary, zone contributions, ops log, well diagram — and the introduction, observations and interpretation, drafted from the computed table and never from figures it was not given. Your analyst edits and signs it before anything leaves.
—The boundary
The measurement is auditable arithmetic. The AI does everything around it.
Writing the interpretation, explaining a refusal in the operator’s language, and answering their questions about the delivered report — the day of assembly, not the judgement. Everything the AI touches is downstream of a number somebody can check, and it is told it may restate those numbers and never derive one.
If a number on your report had come from a language model, we would have to tell you. None of them did.
—The operator
The engineer who reads the report is the reason you win the next job.
A lift engineer wants three things out of a gas lift survey: which valve is actually injecting, where the gas is going, and whether the tubing has a hole. They get all three in a portal under your name, with the charts live rather than flattened into a picture of a chart.
They can ask it questions
Why was valve 4 called open? Is the well multipointing? What is my injection depth? The assistant answers from the recorded reason, declines figures the report does not hold, and will not overturn a call your analyst made.
They can see what was moved
Shift and stretch are recorded per pass, so a depth that was adjusted says so and by how much. Nothing about the alignment is silent.
It is allowed to disagree with them
A downhole total that will not reconcile with the allocated surface split is reported as exactly that, rather than quietly scaled until it agrees. That disagreement is usually the most valuable line in the report.
Your name is on every page
White label throughout — your logo, your colours, your disclaimer. The operator sees your report. They never see us.
—Coverage
The report is chosen from what the file carries.
A caliper log produces a casing inspection. A log carrying QWZI and QWZT produces an injection profile. A flowmeter and a set of valve depths produce a gas lift survey. You pick the passes; the report builds itself around what the file actually holds, and says what it is missing when a page cannot be made. How far the interpretation goes differs by survey type — gas lift is the furthest along, which is why the demo is one.
- GLV optimisation / flowing gradient
- Bottom hole pressure & temperature
- Injection profiles — spinner or tracer
- Step rate & injectivity tests
- Multi-finger caliper / casing inspection
- Dip-in pressure–temperature surveys
—Total depth
Bring a survey where you already know the answer.
There is no sign-up button on purpose. Workspaces are opened by hand, and we would rather run one of your own surveys first — ideally one you have already interpreted, so you can see where we agree with you and where we do not. The second is the more useful half, and it is the fastest way to find out whether this is worth your time.