Runink TIDE Pricing
You choose where Runink TIDE runs: on Runink's shared machines, priced per person; in your own cloud account, priced per node; or on your own premises, priced with you. There is no second bill: TIDE never takes a percentage of anything it finds, fixes or ships, because an audit record or a deploy pipeline is not the kind of thing a success fee should attach to.
That is the whole shape of it. The bill follows the seats or nodes you run, not what TIDE finds, fixes or ships, so a platform team that puts TIDE to heavy use does not open a new line item that grows with it. What the console actually does comes first, though, because that is the part worth arguing about.
The console
One Console, Five Parts
These are the five parts the TIDE console is built from. Each one answers one question, and each opens onto more detail in the TIDE paper.
They are grouped in the order the console's own menu uses them — the order a person goes to when they have the question, not an order invented for this page.
Overview
Is anything wrong, and where do I go next?
-
Needs attention, worst first
Headline readings, then a Needs attention list, worst first, each item linking to the page that deals with it.
-
Ask TIDE in plain words
Ask TIDE, the assistant, does a task in plain words and shows every step it took.
-
Not known is never shown as a zero
A figure the console could not read is shown as not known, never drawn as a zero, so nobody decides on a reading nobody took.
DevEx
Is the platform running, and did our changes ship?
-
GitOps: does it match what was written down?
GitOps shows whether what is running matches what was written down.
-
Deploy lineage
Deploy lineage shows whether a change reached production.
-
Audit chain, verified on demand
Press Verify now on the Audit chain and it names the first record that was altered.
DataEx
What may our models and agents do, and can we trust their work?
-
Model cards, checked against what's live
Model cards list each model's source, version, licence and test evidence, checked against what is live.
-
Autonomy is a setting, never a default
Autonomy is a setting you choose for each kind of action, and every kind starts with a person approving each act.
-
A second opinion, on the record
Where a finding matters, an independent assessor reads the evidence first. It can agree, disagree or say it could not judge.
Intelligence
What does our data hold, and where do money or controls slip?
-
Three dashboards, one set of figures
Dashboards for the analyst, the finance lead and the programme office, built on one set of figures.
-
Computed, or shown absent with the reason
Data quality, capital spending, business rules and lineage, each computed from your own records or shown as absent with the reason.
-
Resolve: an inventory read, not remembered
Resolve maps which systems hold which data and how they connect, when an administrator asks. It keeps structure and counts only, never a value, and marks each link as declared, inferred or a guess.
FORGE (PREVIEW)
How does a written brief become a working application?
-
Describe it, the model proposes the steps
Describe what you want in plain words. The company's own model proposes the steps on a canvas.
-
Nothing filed until a person approves
An approved brief goes, word for word, to coding agents as a work item. Nothing is filed until a person approves it.
-
The brief stays with your own model
The brief stays with the model you run yourself, from the first draft to the working application; we are confirming there is no other path out before promising more than that.
The three licences below differ on one question: where TIDE runs. There is no second question about a percentage of anything TIDE finds, fixes or ships — that question does not exist here.
LITE LICENCE
FOR TEAMS OF 1 TO 9 PEOPLE
PER PERSON, PER MONTH
PER PERSON, PER MONTH
- RUNS ON RUNINK'S SHARED MACHINES, TAKING ITS TURN ALONGSIDE OTHER CUSTOMERS' WORK
- ALL FIVE CONSOLE PARTS: OVERVIEW, DEVEX, DATAEX, INTELLIGENCE, FORGE (PREVIEW)
- MORE UNITS AS YOU GO, OR BOUGHT AHEAD: 10% LESS FOR A MONTH, 20% LESS FOR A YEAR
- UP TO 900 UNITS PER PERSON PER HOUR; WORK PAST THAT PACE WAITS ITS TURN
DEDICATED LICENCE
PRICED PER NODE
A MONTH FOR THE FIRST NODE, THEN $149 FOR EACH EXTRA NODE, ON A YEARLY CONTRACT
ON YOUR OWN CLOUD
- RUNNERS IN YOUR OWN GOOGLE CLOUD PROJECT, DATABRICKS OR SNOWFLAKE ACCOUNT
- ALL FIVE CONSOLE PARTS: OVERVIEW, DEVEX, DATAEX, INTELLIGENCE, FORGE (PREVIEW)
- UNLIMITED COMPUTE UNITS; YOUR CLOUD BILLS YOU FOR THE COMPUTE DIRECTLY
- YOUR OWN WEB ADDRESS
ENTERPRISE LICENCE
FOR RUNNING IT WHERE YOU CHOOSE
PRICED WITH YOU
WHERE YOU RUN IT
- RUNNERS ON YOUR OWN SERVERS, INCLUDING SITES KEPT OFF THE NETWORK
- OR IN YOUR OWN CLOUD ACCOUNT
- UNLIMITED COMPUTE UNITS
- A FULL RECORD OF WHO DID WHAT, AND WHEN
Questions Before You Sign
What the console does and who decides it, then what you are charged for.
We run more software than any one person understands. Will TIDE touch that?
That is the problem TIDE is built for. Most companies run applications from different teams, vendors and years, each with its own way of being started, watched and changed, and its own idea of who may touch the company’s information.
TIDE gives every application the same foundations once: somewhere to run, one way to prove who you are, a way to reach company data, one place to look, and a way to get changes made. The five groups above are that one place to look, and each opens onto the console page that answers the question in its own heading.
What does TIDE need from us before it can do any of this?
Not much to start. One downloaded file and one command bring the whole platform up on a workstation, so you can open Overview and see which readings say they have not been measured, open the Harness and read a remedy, and open the Audit chain and press Verify now — all before anything is installed on machines you own.
Adding one real data source, and putting it on machines you control, are the next steps, each at your own pace. What it needs from your side to begin is a named owner and a decision about who may change data connections and who may arm agents. Everything else can be decided as you go.
Who approves what TIDE's helpers propose?
A person, every time. The automated helpers propose; nothing is filed, changed or published until someone with the right to decide says yes. FORGE will propose the steps to build something on a canvas, and an approved brief only then goes to coding agents as a work item. The Harness takes an action only after a person confirms it, and it is written to the Audit chain before it happens.
That holds however small the action looks. What arrives is always a proposed action, and an action nobody approves is an action that has not been taken.
What does it do when it cannot tell?
It says so, rather than answering anyway. A figure the console could not read is shown as not known, never drawn as a zero, so nobody decides on a reading nobody took. An assessment that could not be completed says so, and says that this is not the same as a clean result.
That matters more than it sounds. The failure this replaces is a plausible-looking answer that nobody downstream could tell apart from a real one.
How do we know the record has not been altered?
Open the Audit chain and press Verify now. The console walks every record and names the first one that was edited, removed, inserted or moved, rather than asking you to trust that nothing changed.
The same page is where GitOps shows whether what is running matches what was written down, and Deploy lineage shows whether a given change reached production — so “what actually shipped” is a reading, not a memory.
Where does our data go?
To the machines your licence names, and it stays there: Runink’s own shared machines on Lite, runners in your own cloud account on Dedicated, your own servers on Enterprise. Nothing is sent to an outside AI service, and nothing is sent out to be trained on — there is no account with an outside model provider for it to be sent to.
Deleting a document removes its influence completely, because nothing is left behind in the model. That is the answer a security review asks for before it will let a platform hold the company’s information at all.
What am I actually paying for?
It depends on where TIDE runs. On Lite, seats: a seat is one person who uses Runink TIDE, and each comes with Compute Units on Runink’s shared machines. On Dedicated, nodes: $249 a month for the first node and $149 for each one after it, with runners in your own cloud account and no limit on units. Your cloud bills you for that compute directly, and Runink adds 1% of what those runners cost. Enterprise runs on your own premises or in your own cloud, priced with you.
You are not billed by the question, the deployment or the finding.
What is a Compute Unit?
It is how Runink counts work on its shared machines. Asking TIDE a question — and the work it does to ground that answer in your own data — draws Compute Units from your allowance.
Every Lite seat carries 18,000 units a month, pooled across the team. On Dedicated and Enterprise the work runs in your own cloud or on your own servers, and units are unlimited.
What happens if we go over the allowance?
Which licence fits us?
Decide where TIDE should run.
Under ten people, the Lite Licence is the fit. It runs on Runink’s shared machines at the lowest price, and it is the only one you can take a month at a time — so an evaluation does not need a year’s commitment. Work there takes its turn alongside other customers’ work.
If TIDE should run in your own Google Cloud project, or your Databricks or Snowflake account, that is Dedicated, priced per node, with unlimited Compute Units. It is taken a year at a time.
If it has to run on your own servers, including sites kept off the network, that is Enterprise, and the conversation starts with where it has to run.
Do we have to sign for a year?
Only for Dedicated and Enterprise. The Lite Licence can be taken month by month at $89 per person, or a year at a time at $75 — the same difference, about 16%, shown by the toggle above.
Dedicated and Enterprise are a year at a time because TIDE runs your platform there: the audit record, the deploy pipeline and the data connections are not something to switch on for a month and off again.
What does Enterprise onboarding and consulting cost?
We already have committed spend with GCP, Snowflake or Databricks. Can we buy TIDE that way?
Yes: that is the Dedicated Licence. Its runners run in your own Google Cloud project, or your Snowflake or Databricks account, on compute you already pay that platform for.
The compute is billed to you directly by the platform rather than provisioned and billed by us, so there is no infrastructure margin of ours stacked on top of theirs: you pay their compute rate, our per-node price, and 1% of what those runners cost. Whatever discount your committed spend gives you is between you and that cloud provider — not a number Runink sets or promises.
Why does TIDE never charge a success fee?
Because TIDE is priced on where it runs, by design, and that was a deliberate choice rather than an oversight. TIDE never takes a percentage of anything it finds, fixes or ships — audit records and deploy pipelines are not the kind of thing a success fee should attach to. There is no reading of “TIDE verified your audit chain” or “TIDE shipped your deploy” that turns into a dollar figure you would want a vendor billing a cut of.
So every licence above is priced by seats or nodes, and by where the work runs — never by what the console found, fixed or shipped.
What does heavier use cost?
On Lite, a team can use up to 900 units per person per hour. Work past that pace waits its turn in line; it is never refused, and the pace itself costs nothing extra. Units beyond the monthly allowance are paid as you go or bought ahead, as above.
On Dedicated, units are unlimited: heavier use shows up on your own cloud bill as the compute it used, plus 1% of that from Runink. On Enterprise it runs where you already run things.
Either way a team that finds heavy use for the assistant, the helpers or FORGE is not billed by the question.