Data Pipeline
Datasets in, SQL transforms on top. Every step is a node you can open and read.
Describe the logic, add the data, ovidens builds the app.
For the reports you produce every month, quarter and year. Turn your logic and spreadsheets into an app, unlocking lasting value from your data through consistent analysis and reliable, repeatable reporting, without starting from scratch each year.
How it works
You bring the logic and the data.ovidens builds the rest.
Your own AI
Bring the AI your team already uses.
Connect your Claude, ChatGPT or any MCP-compatible AI. It acts on your behalf, only within the permissions you and your admin set, and uses no ovidens AI credits.
Audit trail
Calculating the number takes an afternoon.Defending it takes a year.
ovidens keeps the rule, the version and the audit trail behind every figure, so the answer is there if someone asks.
Draft in private
Change a rule or a query in a draft. Viewers keep seeing the published version until you publish.
Fully versioned
Each run records the version it executed and who started it: a person, Vera or a connected system.
Approve & lock
A reporting period can be approved only when every required figure is in and no blocking check fails. Approving freezes the values and publishes one immutable version.
Trace the figure
Follow a result back through the pipeline: the inputs that fed it, the transform that computed it, and the run and version it came from.
No hidden calculation.
Every action, by a person or by Vera, is written to the audit log. Reopening an approved period keeps its version in history.
Who it is for
Three people, one app.Each works at their own depth.
The executive
You approve it. It has to hold.
Every figure keeps the run and version it came from.
The business team
You own the logic, not the SQL.
Describe the rule. AI writes the SQL.
The technical user
Every detail. Nothing hidden.
Raw SQL and the full pipeline, fully editable. Every run and every version, kept.
Products
Three ways data reaches a report.An app for each,versioned the same way.
Bring the data you have. Shape it into the report.
Upload the exports you already have and describe what the report should show. AI builds the pipeline, writes the knowledge base and puts the Visualisation on top, and you can read every step.
Knowledge Base
Every source and transform has a page, so the definition travels with the number.
Vera · AI analyst
You describe the rule. Vera asks back instead of guessing.
Visualisation
Charts, pivots and tables your team edits in place, on top of the pipeline.
Why not the tool you have
You already have tools for numbers.ovidens keeps the reason behind each one.
ovidens is where a recurring report is built once and run again every period. Each figure keeps the rule, the run and the version behind it.
| Where it happens today | In ovidens | |
|---|---|---|
| A spreadsheet | Next quarter's formula overwrites this quarter's, and the file cannot tell you who changed it or why. | The rule is a step in a pipeline. A change is published as a new version, and the change is audited. |
| A BI tool | It draws the chart. The calculation happened somewhere else, before the data arrived. | The data arrives, is calculated and is shown in one app, and all three are versioned together. |
| A data warehouse | It is built for engineers and for queries that answer in milliseconds. A quarterly report needs neither. | The business team describes the report, and the technical user can still read every query. |
A spreadsheet
- Where it happens today
- Next quarter's formula overwrites this quarter's, and the file cannot tell you who changed it or why.
- In ovidens
- The rule is a step in a pipeline. A change is published as a new version, and the change is audited.
A BI tool
- Where it happens today
- It draws the chart. The calculation happened somewhere else, before the data arrived.
- In ovidens
- The data arrives, is calculated and is shown in one app, and all three are versioned together.
A data warehouse
- Where it happens today
- It is built for engineers and for queries that answer in milliseconds. A quarterly report needs neither.
- In ovidens
- The business team describes the report, and the technical user can still read every query.
FAQ
Questions before you start.
A place to build a recurring report once and run it again every period. AI turns your rules and spreadsheets into an app, and every figure keeps the rule and the version it came from.
You do not rebuild the report. Add the new period's data and run it again, and the run is filed under the published version. A change to a rule is made in a private draft and published as the next version.
No. You describe the rule and the AI writes the SQL, and it asks when a rule is unclear. Every query stays readable, so someone who knows SQL can check it.
Yes. A survey app gives you one link anyone can answer without an account. In an indicator app, each reporting unit's figures come from the people responsible for it.
On AWS in Frankfurt (eu-central-1). Each workspace has its own database schema.
Only what the person who asked it may do, checked again on every request. An admin decides, per member, what the AI may do on their behalf.
Yes. You can download an app as a portable bundle, and a connected system can pull a run's results as CSV through the Automation API.
Sign up, name your workspace and upload a file. Billing details are optional at sign-up.
Docs
Read how it works before you start.
The user docs cover the three apps, connecting your own AI, the Automation API and how your data is handled. In English and German.
Next quarter, run it again.
Build the report once. Every period after that, add the new data, run it and publish the new version.