LowestCode
AI NoCode & LowCode Office and Business Operating System
by
The LowestCode Quest:
One AI Agent Guided by One Human Creates an Office and Business Suite in Six Months
Much of what you see runs today, the rest the Quest will show: a range like a large vendor's, possible because LowestCode is one system, not many products.
-
The workplace: several windows of one system side by side, in three themes
-
The Quest – updated every month
-
Import and export with Office tools
-
PMS: projects and workflows, declared and checked
-
The rollout and contribution chain
-
Energy and server hardware per 1,000 users
-
Innovation and cooperation
Click to jump to a section: the Quest, where our projects stand, what we modernize for customers, how your team works with it, the platform behind it, the workplace, the energy and server comparison, innovation and cooperation, and the summary.
The Quest.
One agent builds.
LowestCode controls.
Six months, October 2026 to March 2027, with monthly KPIs as public, verifiable proof.
One AI agent creates DMS, CRM, HRM, FIN, PMS, ERP, Insurance, Office, Website, Design, Communication, Statistics, Sim-Test, and more on LowestCode AI. The LowestCode Communication Protocol (LCP) controls coding and execution in every phase.
We don't write code anymore
Since September 2026, 100 % of our development runs through AI agents, in harmony with the system: LowestCode conditions the agent, answers every coding step with live feedback, and its security and controls step in when needed. We set direction and review. Users and developers will create applications the same way. Today one person does this work: the founder sets direction and reviews, the agent codes. The platform under it is ten years of development, with much input from real projects.
What the Quest will prove
Safe and controlled AI-only coding is reality and delivers.
One system for all common Office and business tasks.
Qualified AI / human error statistics: once obvious errors are handled, a remaining error is as often a definition, a preparation, or a human error as the AI's.
80 to 99 % less energy and server hardware, about 90 % in the middle of that range, compared with Dynamics and classic ERP stacks.
Less effort and energy for AI application creation, through LowestCode automation and self-learning.
10 to 30 times faster for core database and cloud operations, compared with the same stacks.
AI and cloud security with data protection, each designed for its EU rules: the EU AI Act for the AI, GDPR and the EU cybersecurity rules for the cloud.
A better user interface with multiple windows and built-in AI support for documents and coding.
21 application languages from the start.
Where our projects stand
- DMS
- CRM
- HRM
- FIN
- PMS
- ERP
- E-Commerce
- Logistics
- Production
- Planning
- Insurance
- Office
- Website
- Design
- Communication
- Statistics
- Sim-Test
- Python-App
- Desktop-App
- Mobile-App
- Native-Apps
- Data-Warehouse
- MCP
| Project | September 2026 | October 2026 | Start | Description | ||
|---|---|---|---|---|---|---|
| Complete | Effort | Complete | Effort | |||
| LowestCode | ✅ 80 % | 7 years | ✅ 80 % | 7 years | 2016 | AI NoCode & LowCode Office and Business Operating System |
| DMS | ✅ 70 % | 1 year | ✅ 70 % | 1 year | 2020 | Document Management System with Nested, Multilingual Documents and Version Control |
| IMS Art Insurance | ✅ 90 % | 2 years | ✅ 90 % | 2 years | 2021 | Art & Real Estate, Excel import, in production since 2024 |
| IMS Art API | 🟦 90 % | 4 months | 🟦 90 % | 4 months | 2024 | Art & Real Estate API, approved and ready 2026 |
| AI Deep Dive | ✅ 80 % | 1 year | ✅ 80 % | 1 year | 2025 | LowestCode deep dive AI integration: AI-for-AI documentation and controls; the agent builds from it and explains the system to people |
| IMS Car Insurance | 🟨 70 % | 3 months | 🟨 70 % | 3 months | 2026 | Car Fleet Insurance, Excel import, in development / test |
| Sim-Test | ✅ 50 % | 6 days | ✅ 50 % | 6 days | 2026-04 | Simulation & Test Data Generator: Copy / Generate / Anonymize / Import |
| LCP-Chat | 🟨 40 % | 5 days | 🟨 40 % | 5 days | 2026-08 | LowestCode Communication Protocol: Document & LCP Request AI Chat; System Chat in preparation |
| Office-Text | 🟧 20 % | 10 days | 🟧 20 % | 10 days | 2026-08 | LowestCode Office Text Documents with Word & PDF Import/Export |
| Office-Calc | 🟧 10 % | 5 days | 🟧 10 % | 5 days | 2026-08 | LowestCode Office Table Documents with Excel Import/Export |
| PMS | 🟧 5 % | 2 days | 🟧 5 % | 2 days | 2026-09 | Project Management System - prepared for structural integration |
| … | … | … | … | … | … | More projects to start soon |
✅ In production · 🟦 Ready · 🟨 In test · 🟧 In preparation · ❌ Canceled
September 2026 shows the start values of the Quest, October 2026 the current values. Completeness is estimated, effort is the working time spent. A value that did not change since the month before is shown muted.
… The table is open: a planned project gets its row when work on it starts, and the table grows until every planned project is built out.
From proven workflows to maintainable applications Modernize business-critical workflows without losing proven business logic.
We transform insurance processes, enterprise workflows, Excel workbooks, and VBA applications into maintainable software for European cloud and desktop environments. LowestCode begins with the business logic that already creates value and gives it a controlled path forward.
-
Insurance
Products, calculations, documents, and processes developed around real operational needs.
-
Excel & VBA
Valuable workbook logic moved into maintainable, shared applications.
-
Enterprise workflows
Data, forms, rights, documents, and automation brought into one model.
-
European deployment
Cloud, private, and desktop options, designed for EU data protection and conformity.
- Preserve business logic
- Reduce fragmentation
- Control change
- Support long-term use
Power users deliver.
IT experts guide.
Development partners join through one governed rollout chain.
The LowestCode rollout system lets development partners build independently without interfering with one another, as the integrated AI agent does the coding and keeps it consistent.
- System
- Professionals
- Organization
- User
Contributions travel up the chain in a controlled way. Once we are established on the market, we plan a participation system that rewards lasting contributions.
- User
- Organization
- Professionals
- System
-
Power users meet the internal needs
The people who know the work build the forms, documents, and workflows they need, without code, and make adjustments themselves as the business changes.
-
IT experts guide and control
An IT expert introduces each project, advises and accompanies the power user, checks the result, and gives the final approval. Architecture, security, and data stay in professional hands.
The AI agent does the coding, with the system's expertise built in and its checks behind every step. -
IT experts freed for what usually waits
Relieved of routine application work, IT experts take on the higher-level tasks that normally get neglected: integration, security, and intelligent global solutions and helpers for the whole company.
-
A better-run company
Internal needs are met quickly, professional control stays in place, and expert time goes where it counts: an overall improvement of company management and business development.
- Power users build
- IT experts approve
- Partners contribute
- Management improves
Built from real use, governed as one system LowestCode is what AI needs: one integral, compact system for Office and business applications, and LCP controls every step. The result is controlled agent coding, high coding efficiency, and a clear interface for people.
Productive insurance software has been licensed and used since 2024, with continuing development and integration into established systems. LowestCode preserves proven logic while modernizing how it is operated.
Why LowestCode
What a power user needs without code is what an agent needs. Data, interfaces, documents, rights, translations, and automation remain governed parts of one application system. Embedded in this architecture and guided by LCP, an AI agent implements faster and more consistently than hand coding; people set direction and review, and system rules retain control.
Preserve valuable business rules instead of starting from zero.
Generate consistent applications from governed structures.
Your applications will be created and maintained by an AI agent under LowestCode's control, from the first version through every later change.
Validate and reverse controlled NoCode, LowCode, and AI-assisted changes.
Data carries its validity: every row has a valid-from and a valid-to time, so parameters, taxes, texts, documents and user access are filtered and validated by date, and a change can be timed in advance. Beside that, changes to selected fields are recorded with author and time, and documents keep complete versions.
21 application languages from the start, and each partner can add languages or define dialect variants; right-to-left scripts, time zones, country formats and customer-specific wording are built in.
No license fees for components: every third-party component is free open source. AI agents and cloud hosting are external services with their own terms.
Deploy across European cloud, private environments, and controlled desktop use.
Customer data stays safe in production, where users can still make their own adaptations; deeper development runs only in synthetic test environments.
Java / Kotlin for the system, Python for science
LowestCode runs on Java / Kotlin: strict types and fixed contracts keep the system deterministic. Python complements it where science and data science need it, with its analysis, AI and machine-learning libraries. The Python version is designed as a full cloud version of LowestCode, not a set of tools: it shares the database and the user interface with the Java / Kotlin version, so it can deliver results directly to users or through APIs. A first Python runtime slice exists; the data-science services are planned. PostgreSQL is the standard database; SQLite is planned for mobile, and MS SQL and MySQL serve compatibility tests.
One workplace.
One development IDE.
Many windows.
One shell for any application, the Office suite on its way.
Your windows, ready when you log in
Your windows and their layout are saved and preloaded, ready the moment you log in. Working environments can be defined for each task and guarded by rights. Windows sit side by side, and the same document can be open in many of them, working in sync. The same system is your workplace and your development and test environment, so you build your own applications there and test them on synthetic data before they reach production.
One menu, two views, one toolbar
The navigation is one menu shown two ways, as a side tree or as a top menu. The side menu doubles as the toolbar, with the HTML editor's controls in it, as on the left of the picture. Themes and formats change the look, not the content.
Office on its way
Planned: text documents with Word and PDF import and export, table documents with Excel import and export, presentations with PowerPoint, and notes, tasks, reports, forms and designs, all governed documents of the one foundation, to complement or replace Microsoft Office and Adobe tools. Office is the focus among the Quest's four expansion candidates, starting October 2026.
95 % less energy and server hardware on average Compared with six leading business systems; SAP comes closest.
LowestCode keeps about one core busy for 1,000 users, measured; sized with headroom and a standby, that is four cores, eight with high availability on two sites. Five of the six other systems need far more by their own published sizing rules.
- 1.4 billionrows read by the database in one run
- 30 minutesof full load
- 171,000checks of the answers, none failed
- 9 to 136 msstandard response time: 50 % to 95 % of the requests
- ≈ 5 wattsprocessor power, load generator included
The numbers, their sources and the measurement are on these pages; no competitor product was measured. All product names are trademarks of their respective owners.
A structured innovation portfolio grounded in working software Deterministic AI, internationalization, digital sovereignty, and measurable efficiency.
We have structured a portfolio of 93 innovations and innovation candidates, including 42 system-level architecture innovations (A-category). They rethink how database semantics, documents and interfaces, automation, AI governance, security, and deployment work together.
On this basis we are preparing a research grant application to the German research ministry (BMFTR, formerly BMBF) and are looking for a university partner: companies are funded at 50 to 70 %, universities up to 100 %. The project validates what the working system shows today: energy efficiency, security by design for the EU rules, and full and secure AI integration in application creation and documents. Our thesis: AI needs an integral, compact system with governed controls. LowestCode is built that way, and LCP controls every step.
-
Customers
Modernize one valuable workflow with a practical delivery path.
-
Developers
Build and improve governed applications with inspectable tools.
-
Universities
Study reproducibility, AI control, efficiency, and multilingual systems.
-
Investors and partners
Support customer-led growth and the validation of a reusable platform.
AI providers and cloud hosting
Offer the LowestCode Office and Business Operating System as your own application layer. Your model becomes the agent that creates and maintains every application throughout its life, and LowestCode keeps it on track: it prepares the agent, checks every step, and decides the rights once before anything runs. Every seat adds recurring revenue on top of your models and hosting. Terms are open, from distribution to deeper collaboration, and every cooperation starts small, with a technical review or a pilot.
Implementation partners
Bring LowestCode to your customers: insurance modernization, Excel and VBA migration, and business applications. The AI agent does the coding, and your team guides it.
Self-learning, built to continue
Accepted AI work becomes predefined, reusable modules without training a model, and deterministic controls keep authority, auditability and rollback. A governed step runs 10 to 100 times faster than an AI agent reasoning it through, and more safely: within narrow controls, with the owner and user rights applied to every step. The knowledge lives in the system, not in one head: every working session starts a fresh agent that picks up the system from there, and a partner team joins the same way.
AI-for-AI: one documentation, two uses
LowestCode's documentation is written by the AI agent for the AI agent: compact, structured and complete. It covers the whole system and states for each part what is built and what is planned. That gives it a second use. The same agent explains the system to people in detail: a developer asking how a rule works, a specialist joining a project, a partner team getting started. The answers come from the current documentation, not from a manual that fell behind.
LowestCode in short One integral, compact system for Office and business applications, and LCP controls every step. Many problems solved.
In use and measured today
- ✅Ten years of development, proven in real use: insurance software licensed and in production since 2024.
- 🧠We don't write code anymore: since September 2026, 100 % of our development runs through AI agents, and the LowestCode Communication Protocol (LCP) controls coding and execution in every phase.
- ⚡About one busy core for 1,000 users, measured: in one 44-minute load run the database reads more than 1.4 billion rows, and the device does not even get warm.
- ⚡80 to 99 % less energy and server hardware than five of six leading business systems, from our measured load and their own published sizing rules.
- 🌍21 application languages from the start, and each partner can add languages or define dialect variants; right-to-left scripts, time zones and country formats are built in.
- 🆓No license fees for components: every third-party component is free open source. AI agents and cloud hosting are external services with their own terms.
How it works
- 👥Power users deliver, IT experts guide and approve, and the AI agent does the coding.
- 🏅One rollout and contribution chain: System ⬌ Professionals ⬌ Organization ⬌ User. Partners build independently, and contributions travel back up in a controlled way.
- 🔁Workplace, development IDE and test environment in one: customer data stays safe in production, deeper development runs only in synthetic test environments, and changes can be validated and reversed.
- 🛡️AI and cloud security with data protection, each designed for its EU rules: the EU AI Act for the AI, GDPR and the EU cybersecurity rules for the cloud.
- ♻️Self-learning by design: accepted AI work becomes predefined, reusable modules without training a model; a governed step runs 10 to 100 times faster than an AI agent reasoning it through.
- ❤️AI-for-AI documentation: written by the agent for the agent, covering the whole system and kept up to date with every change; the same agent explains the system to people.
- 🧩Cross-platform by design: Java / Kotlin for the system, Python for science and data science.
- 🗄️PostgreSQL as the standard database, SQLite planned for mobile, MS SQL and MySQL for compatibility tests.
What the Quest will show
- 🚀The Quest: one AI agent, guided by one human, creates an Office and Business Suite in six months, October 2026 to March 2027, with monthly, public, verifiable proof.
- 📄One system for all common Office and business tasks: the Office suite is on its way, to complement or replace Microsoft Office and Adobe tools.
- ⚙️Spontaneous RPA in the same system: the AI agent writes an automation in KotlinScript when it is needed, the owner approves it, and LowestCode controls the coding and decides the rights before every run; an approved automation runs again on demand, without the agent.
- ⚡Designed to run core database and cloud operations 10 to 30 times faster than Dynamics, SAP, and classic ERP stacks; the Quest will measure it.
- 🤝Technology partners welcome now, before sales start at the end of the Quest in March 2027; a university partner is sought for a research grant application.
LowestCode© Dr.-Ing. Raimund Reisacher
This document is available in English only. It is the text we check and verify.
Energy and Server Comparison
© RSaccess
This page compares the server hardware and the power that LowestCode and six leading business systems need for 1,000 users. It shows the numbers, says where each one comes from, and says what the comparison proves and what it does not. State: 1 October 2026.
The result in short
- LowestCode needs 80 to 99 % less energy and server hardware than five of the six systems, about 90 % in the middle of that range.
- Against SAP S/4HANA it needs about 89 % fewer server cores and 72 to 80 % less power. SAP's work per user is small; its hardware comes from the database held in memory and the replicated host that guards it.
- Against the mean values of the six: 95 % less power and 95 % fewer server cores.
- The LowestCode side is measured and then sized for real operation. The other side comes from the providers' own published sizing rules and certified benchmarks. No competitor product was measured.
The numbers
Server cores and power per 1,000 users. Each row is a range: from the smallest setup a product runs on, to a landscape with high availability.
|
Provider
|
Server cores
|
Power
|
Basis
|
|---|---|---|---|
|
LowestCode
|
4–8
|
20–31 W
|
A
|
|
SAP S/4HANA
|
36–76
|
71–154 W
|
B
|
|
Microsoft Dynamics 365 F&O, on-premises
|
200–290
|
600–2,000 W
|
B
|
|
Oracle JD Edwards, E-Business Suite
|
64–240
|
250–800 W
|
C
|
|
Infor LN
|
100–200
|
250–1,000 W
|
C
|
|
Sage X3
|
40–120
|
200–600 W
|
C
|
|
Odoo Community
|
83–200
|
130–500 W
|
B/C
|
- A is measured.
- B is a published sizing rule or certified benchmark of the provider, put on measured server power.
- C is the shape of a typical landscape of that product, or an analogy.
The estimates of the six providers are accurate to about ±30 %. Every row is scaled to 1,000 users and rests on the provider's own published sizing rules, which include headroom.
One load run you can picture
1,000 simulated office users, 44 minutes, one device: 3 minutes idle, 10 minutes ramp-up, 30 minutes under full load, then the ramp-down. In that time the users perform about 141,000 actions; each of the 171,000 answers is checked for its content, and none fails. To serve them, the database reads more than 1.4 billion rows. The processor draws about 5 watts for the whole test, the load generator included. The device does not even get warm.
The full record of the five runs is the load test report.
Where each number comes from
LowestCode. We measured the load: 1,000 users keep 1.2 processor cores busy, server and database together. A real installation needs more than the busy cores, so we sized it. For single operation the machines run at about 40 % load and one standby machine waits: 4 cores and 20 watts per 1,000 users. For high availability the same setup stands on two sites, and either site can carry the full load alone: 8 cores and 31 watts. The measurement is described in the load test report.
The arithmetic, done at 10,000 users: the 12.1 busy cores run on four servers with 8 cores each, at about 40 % load, and a fifth server waits as standby, 40 cores in all. By its published power data, a Dell PowerEdge T150 draws about 21 watts idle and 84 watts at full load. At 40 % load each active server draws about 45 watts and the standby about 21 watts, together about 201 watts: 4 cores and 20 watts per 1,000 users. On two sites, ten such servers, eight of them sharing the load and two waiting, come to 80 cores and about 307 watts: 8 cores and 31 watts per 1,000 users.
SAP S/4HANA. We sized SAP the way SAP sizes an installation, with its own published rules. SAP's Quick Sizer counts users by activity: a medium user works one step every 30 seconds, a high user one every 10 seconds; our users work one every 15 seconds. SAP sizes the processors for an average load of one third. For 1,000 users that gives an application server of 4 to 6 cores. The database of S/4HANA, SAP HANA, holds its data in memory, and SAP certifies productive HANA servers only from 256 GB of memory with 16 cores, as the published lists of AWS, Google Cloud and Microsoft Azure show; for availability a second server of the same size keeps a replicated copy. That memory has its own price: server memory has become very expensive, and its supply is uncertain, so a database held in memory costs more than its watts show. That makes about 36 to 38 cores and 71 to 77 watts for single operation, on the published power curve of the Lenovo SR650 V3, the server of SAP's own certified benchmark; with high availability on two sites 72 to 76 cores and 142 to 154 watts. Two caveats: the processor share per user follows SAP's classic factors, and S/4HANA's own factors are not published and probably higher; the database server is the smallest certified size, while a real database of 1,000 users may need more memory. Both make SAP's row a lower estimate. SAP's certified benchmark itself, 65,208 users on one 120-core server, ran SAP ERP 6.0 on IBM Db2, not S/4HANA: it shows how efficient SAP's order-processing core is in a best case, not how large an installation is.
Microsoft Dynamics 365 Finance and Operations, on-premises. Microsoft publishes a sizing rule in users per core and a minimum topology with spare nodes and a paired database server. Applied to 1,000 users this gives 200 to 290 cores.
Oracle JD Edwards and E-Business Suite. We found no per-user rule, so the range follows the shape of a typical landscape: database and application hosts, from a single machine to high availability. This is the least certain kind of estimate.
Infor LN. Infor's minimum hardware guide names 10 cores for 100 concurrent users. Scaled to 1,000 users that is 100 cores, and 200 with high availability.
Sage X3. Sage publishes a sizing rule for its web tier. The application, database and print hosts of a usual installation come on top.
Odoo Community. Odoo publishes how many concurrent users one processor serves. The database comes on top, from a single machine to a pair.
The power figures put these cores on servers with published, measured power data.
What the numbers say, and what they do not
They say that LowestCode needs far fewer and far smaller machines than five of the six systems: against their documented ranges it is 80 to 99 % less energy and server hardware. Against SAP S/4HANA the difference is smaller, about 89 % fewer server cores and 72 to 80 % less power: SAP comes closest, because its hardware is set by the database held in memory rather than by the work of its users. Against the mean values of the six, LowestCode needs 95 % less power and 95 % fewer server cores.
|
Provider
|
Mean cores
|
Times LowestCode
|
Mean power
|
Times LowestCode
|
|---|---|---|---|---|
|
LowestCode
|
6
|
25.5 W
|
||
|
SAP S/4HANA
|
56
|
9×
|
112.5 W
|
4.4×
|
|
Microsoft Dynamics 365 F&O, on-premises
|
245
|
41×
|
1,300 W
|
51×
|
|
Oracle JD Edwards, E-Business Suite
|
152
|
25×
|
525 W
|
21×
|
|
Infor LN
|
150
|
25×
|
625 W
|
25×
|
|
Sage X3
|
80
|
13×
|
400 W
|
16×
|
|
Odoo Community
|
141.5
|
24×
|
315 W
|
12×
|
|
Mean of the six
|
137
|
23×
|
546 W
|
21×
|
They do not say that anyone measured a competitor. The six rows are estimates from the providers' own documents. No other system was measured on this workload.
Where the difference comes from. The work itself is small. SAP's certified benchmark, a best case on one tuned standard process of SAP ERP 6.0, needs between one and two busy cores for 1,000 of its benchmark users, the same class as our measured 1.2 cores. The work of 1,000 office users is a few watts for a well-built system. The energy goes into the machines a product needs in order to run at all: its separate tiers, its minimum sizes, its spare capacity, and for SAP the database held in memory. LowestCode runs as one small application with its database, and that is where the saving is.
Calculated first, then measured. The result is no surprise. Months before the load tests we compared LowestCode with a conventional stack layer by layer: application, database, cache, search, document storage, queue and monitoring servers. That calculation gave 80 to 96 % less energy and server hardware, about 90 % in the middle. The measured load, sized for operation, gives 95 % against the mean of the six documented systems: the upper end of what we calculated.
What is included. The savings come from one model, not from leaving things out. Every table carries a validity interval, soft deletion, an audit trail and rights on three levels; every text field is multilingual; times are stored in UTC and shown in the user's time zone; tenants are isolated by schema and every environment by its operation mode. Forms and windows are prepared once and loaded ready, instead of being assembled on every request.
Sources
- SAP sizing (sap.com/about/benchmark/sizing.html), Quick Sizer user classes and SAPS definition
- SAP Standard Application Benchmark, SD definition and certification Cert23002 (Lenovo Press lp1685)
- SAP HANA certified instances: AWS (docs.aws.amazon.com/sap/latest/general/sap-hana-aws-ec2.html), Google Cloud (docs.cloud.google.com/sap/docs/certifications-sap-hana), Microsoft Azure (learn.microsoft.com/en-us/azure/sap/workloads/hana-vm-premium-ssd-v1)
- SPECpower_ssj2008 results 20230619-01282, 20240328-01390 and 20230829-01308
- Apple support article 103253
- Microsoft Learn, hardware sizing requirements for on-premises environments
- Odoo 18.0 deployment documentation
- Infor LN 10.3 minimum hardware requirements
- Sage X3 Syracuse sizing recommendations
All product names are trademarks of their respective owners.