UMAY OS

Models change.Experience endures.Decisions remain human.

An isolated Linux workspace for every employee. SWAPP, AI-Scientist and other agents run here; operations become records, and valuable experience improves local models.

Target architecture · Containers per employee · Local operation · Architecture developed with powerful models

Umay Ana opening animation

Plays with sound; the level is the file's own.

More than an answer.A way of working.

Turn human experience into a lasting company capability.

  1. An isolated workspace for every employee.

    A Linux container starts for each employee; the SWAPP web application, AI-Scientist and other agents run within this isolated scope. Sessions, files and access permissions stay separate; company knowledge remains private.

  2. Record every operation. Retain valuable experience.

    Human, agent and service operations in the workspace are logged with goals, actions, outcomes, failures and corrections. Records persist independently of containers. Only permitted examples with verified value enter learning.

  3. Develop architecture with powerful models. Operate locally.

    Architecture changes are prepared with powerful large language models: decision rationale, code changes, tests and rollback plans are reviewed together. People approve versioned releases. Daily agents run open-weight models locally.

  4. Measure mastery of SWAPP.

    Recognizing screens is not enough. An agent must demonstrate finding assets, interpreting data and completing permitted workflows with independent outcome verification.

  5. Keep authority human.

    Models do not grant permission. Tools and access are scoped to the employee and task. People decide, accept learning records and approve new versions for use.

  6. Start with one specialist. Welcome more.

    AI-Scientist is the first specialist; other local agents can join through shared task, permission, logging and result contracts. Every specialist runs in an isolated scope; scientific experiments retain reproducible data, methods and results.

  7. Act with skills. Retrieve with RAG. Improve with fine-tuning.

    Selected records become tested skills, sourced RAG knowledge and fine-tuning candidates for local models. LoRA/QLoRA is evaluated only for compatible models. Measure each candidate’s contribution; raw logs are not training data by default.

  8. Measure improvement and cost together.

    Track time and resource use for every model call, experiment and training job. Evaluate quality, latency and cost together; unmeasured savings or charges are never treated as zero.

A good answer is a beginning.Verified knowledge is a traceable record.

A workspace for each employee.Shared learning.

AOS Core manages an isolated Linux container scope for each employee. Each workspace runs the SWAPP web application, AI-Scientist and other specialist agents. Local model services and persistent records remain in the company environment. Integrated acceptance of this target architecture is still pending.

Three foundations, clear responsibilities.

AOS

Core and execution

Local models, task policy and tool execution. The source foundation for a company-adapted Core.

SWAPP

The human workspace

Assets, time series and digital-twin context. The public presentation is accessible; application code and company content are private.

AI-Scientist

Experiments and evaluation

The scientific specialist foundation for clustering, anomaly detection and reproducible experiments.

Operation

Company environment · Linux host

Isolated Linux container workspace per employee

Separate sessions, files, network access and resource quotas. The same template is provisioned independently for every employee.

Employee session · purpose and permission

UMAY OS Core · AOS

System 1 · fast, bounded action selection

System 2 · planning, explanation and recovery

AI-Scientist · first specialist

Isolated experiment jobs scoped to the employee

SWAPP · web application

Frontend + backend · an instance per employee

Other agents

Task-scoped tools and permissions

Persistent company services outside containers

Local model services
S1 / S2 / research · separate access and context
Persistent logs and learning store
All operations → value selection → skills / RAG / fine-tuning

A model is a compute resource. An agent is an execution role with defined boundaries.

An employee workspace consists of one container or service containers within the same isolation scope. SWAPP, Core and specialists use separate process permissions; Scientist experiments run in additional sandboxes. Core is the sole automation owner of each browser session. Records and approved learning artifacts survive container recreation. Scientist’s existing broker remains the single shared-GPU allocation authority; employee containers do not create separate GPU schedulers.

System 1 and System 2: separate roles, separate learning.

S1 · Operator

From permitted options to action.

The AOS source baseline uses Mapika/decider-2b. It selects from permitted options using a fresh observation, rather than acting as a free-text chat model. Decision data, training objectives and adapters are specific to S1.

S2 · Supervisor

From goals to plans and explanations.

The AOS source baseline uses Ternary Bonsai-2 27B for planning, sourced explanation and recovery. S2 has separate data and adapter profiles; it is distinct from the closed-source teacher line.

Unsloth · training line

Model-specific LoRA / QLoRA.

Each role can move to current compatible models. Unsloth training uses supported base weights and role-specific data formats. Adapter loading and benefit in the target inference engine require separate validation.

These model names identify the inspected AOS baseline, not permanent product choices. Unsloth compatibility is not assumed for Decider or ternary/GGUF Bonsai. Validate LoRA and 4-bit QLoRA support per model; unsupported models require a separate backend or an evaluated alternative. Scientist research profiles also differ from the Core S1/S2 protocol.

An open system for new local specialists.

AI-Scientist is the first specialist. Other locally running specialist agents can join through versioned capability registration, task/result contracts, scoped tools and data, resource quotas, cancellation and operation logs. Every specialist follows workspace isolation and human oversight.

The new registered-agent runner in AOS provides a source foundation for this extension. A stable general plugin API, compatibility with every specialist and multi-employee UMAY acceptance are still pending. Open here means extensible; project licensing is a separate decision.

Architecture changes with powerful large language models.

  1. 01Claude / OpenAI · powerful development model
  2. 02Architecture decision · code changes · rollback plan
  3. 03Isolated tests · quality and cost
  4. 04Human review and approval
  5. 05Versioned deployment · local operation

Architecture and contract changes are prepared with powerful large language models in a separate development workspace; rationale, affected components and test evidence are recorded together. Daily local agents do not change the architecture during operation. Company records are not automatically sent to external providers. Model identity and cost are tracked; local operation should continue without teacher access.

Contracts before integration. Evidence before promotion.

  1. 01Employee containers + application contract
  2. 02Persistent logs for all operations
  3. 03Valuable records → skills / RAG / fine-tuning
  4. 04Independent quality and cost evaluation
  5. 05Human promotion · rollback

First test isolation, record integrity and restarts with two synthetic employee workspaces. Once private SWAPP code is delivered, validate real workflows in the same scope. Measure the learning value of selected records separately; architecture changes are prepared with powerful models and pass tests and human approval.

Two separate architectural workspaces for candidate production and reviewed modules. Conceptual composition.
Two workspaces, two roles. Review stands between production and use. Conceptual artwork.

First specialist: Scientist.

First task: from digital twin to scientific experiment.

AI-Scientist will run as the first specialist in each employee’s Linux container workspace. Clustering and anomaly experiments on permitted SWAPP digital-twin data will run in separate sandboxes; every trial, failure and result will link to the shared operation trace. Research models will be evaluated separately from Core S1/S2 roles. The new source includes CPU experiments and selected experiment memory; LLM/GPU calls are disabled in the default field-lab profile. Integration into UMAY employee workspaces requires separate acceptance.

The planned first research pipeline

01 · Digital-twin data

Versioned data with asset, time, units and quality. Independent test splits by time and asset.

02 · Operating regimes

Explore similar behavior through clustering. Review clusters with domain knowledge; cluster labels are not ground truth.

03 · Anomalies and experiments

Compare against simple baselines; measure false alerts and stability. An anomaly score is neither a fault diagnosis nor a control command.

Illustration · six synthetic points · not an ML model

1 / 6 · Question

How did past vibration change?

Consider six days of history for an anonymous rotating asset. The aim is to compare possible explanations, not to intervene in a control system.

2 / 6 · Data

Permitted data. Explicit provenance.

Six entirely synthetic samples. Time: 1–6 September 2026, daily at 12:00 UTC. Unit: mm/s RMS. These values do not come from a real facility or device.

3 / 6 · Calculation

Deterministic calculation. Visible method.

vibration-window-mean v1 compares arithmetic means of the first and last three samples. Percentage change = (last mean / first mean − 1) × 100. The same data produces the same result.

4 / 6 · Hypotheses

One finding. Several explanations.

Changes in load or speed, measurement conditions, or mechanical condition could explain the difference. Vibration growth alone cannot distinguish them or diagnose a fault.

5 / 6 · Answer

A sourced Advisory Packet.

The packet carries the question, source version, method, supporting and contradicting evidence, alternatives and missing data. Uncertainty remains visible.

6 / 6 · Review

Lasting Investigation. Human decision.

A human reviews the evidence and identifies the next data request. An asset-linked Investigation is intended to retain sources and review history. This page runs no record service.

Vibration · mm/s RMS
Vibration · mm/s RMSIllustration · six synthetic points · not an ML model. 2026-09-01T12:00:00Z — 2026-09-06T12:00:00Z. 2, 2.2, 2.1, 2.5, 3.2, 3.4 mm/s RMS.0123452.001 Sep2.202 Sep2.103 Sep2.504 Sep3.205 Sep3.406 SepVibration · mm/s RMSIllustration · six synthetic points · not an ML model. 2026-09-01T12:00:00Z — 2026-09-06T12:00:00Z. 2, 2.2, 2.1, 2.5, 3.2, 3.4 mm/s RMS.0123452.001 Sep2.202 Sep2.103 Sep2.504 Sep3.205 Sep3.406 Sep

2026 · 12:00 UTC · vibration-window-mean v1

Deterministic calculation. Visible method.

First three days
2.10 mm/s
Last three days
3.03 mm/s
Change
+44.4%

This change is not a fault diagnosis.

Source and method

synthetic-vibration-v1.csv

vibration-window-mean v1
2026-09-01 → 2026-09-06 · UTC

Alternative explanation

Load or measurement conditions may have changed. Mechanical condition cannot yet be distinguished.

Missing evidence

Speed, load, sensor calibration and maintenance records are absent.

Chart data table
Time (UTC)Vibration (mm/s RMS)
2026-09-01T12:00:00Z2.0
2026-09-02T12:00:00Z2.2
2026-09-03T12:00:00Z2.1
2026-09-04T12:00:00Z2.5
2026-09-05T12:00:00Z3.2
2026-09-06T12:00:00Z3.4

Read and advise. Writing to control systems, setpoints, start/stop and alarm-state changes are out of scope.

Models change. Knowledge keeps its history.

Log SWAPP use, agent steps, model and tool calls, experiments, failures and human corrections within each employee workspace. Select records with verified outcomes and reusable value; review them for skills, RAG and local model training.

Transparent archive layers representing knowledge provenance and review history. Conceptual composition.

RAG

Retrieve knowledge; keep weights unchanged.

Reviewed documents and workflow knowledge form a private vector dataset with access scope and source versions. Agents use relevant knowledge in answers or task context.

Skill

Make workflows repeatable.

Version a SWAPP workflow or agent method with preconditions, steps, authority, expected outcomes and tests. A record becomes a skill after its repeatable contribution is verified.

Fine-tuning · LoRA / QLoRA

Separate, model-bound adapters for S1 and S2.

Reviewed examples are converted to the role’s data format. Choose LoRA or a compatible 4-bit QLoRA path for Unsloth-supported models. Evaluate the adapter against the same base on independent tasks before human-approved use.

  1. Log all workspace operations
  2. Select and review valuable records
  3. Skill / RAG / fine-tuning candidate
  4. Independent local-model evaluation
  5. Human approval, version and rollback

Logging an operation does not grant permission to train on every payload. Remove secrets and unnecessary personal data; failures and corrections can also be valuable. RAG does not change weights. Fine-tuning candidates enter local operation only after independent evaluation and human approval.

Where do we stand today?

Existing in public sources

The new pushes were reviewed at source level. Execution results below are dated upstream acceptance reports, not UMAY runtime tests rerun on this site.

  • AOS: registered specialist runner, durable task events and isolated project lifecycle. Limited real-experiment acceptance is reported for the Scientist CPU adapter.
  • AI-Scientist: v0.1.0 product delivery / 0.46.0 laboratory; CPU experiments, independent Scorer/Referee and explicitly selected experiment memory.

UMAY architecture decision

  • An isolated Linux container workspace per employee: SWAPP + AOS Core + agents.
  • AI-Scientist first; other specialists use the same task and logging contracts.
  • Log all operations; turn valuable records into skills, RAG and local fine-tuning.
  • Powerful large language models, tests and human approval for architecture changes.

Integration and acceptance pending

  • Employee container workspaces, new local specialist acceptance and persistent log integrity.
  • Private SWAPP code delivery and real workflow acceptance.
  • Complete operation logging, value selection and measured local-model improvement.
  • A compatible AOS–Scientist source/contract pair and real shared-GPU acceptance under one authority.
  1. Foundations and contracts

    Employee container template, access boundaries and persistent operation ledger.

  2. SWAPP and first experiment

    From synthetic workflows to the real application; a digital-twin experiment setup.

  3. Learning and acceptance

    Valuable records → skills / RAG / fine-tuning; independent local-model comparisons.

  4. Continuous improvement

    Architecture developed with powerful models, human approval and reversible releases.

Every improvement has a quality result and a cost record.

Track teacher calls, coding, experiments, GPU training and local inference separately. Keep budget estimates, measured usage and reconciled invoices distinct. Include failed jobs and retries in costs.

A starting point for Claude and ChatGPT

Employee container architecture, operation records, value selection, local learning and changes with powerful models; contracts, acceptance criteria and a handoff prompt in one pack.

Download implementation pack · Markdown (Turkish)

Project licenses for these repositories have not been selected; public access does not grant commercial reuse permission. Verify controlled forks and usage scope against source and model licenses. This manifesto, plan and example do not prove a working company integration.

Intelligence can change.Responsibility endures.

Turn verified experience from each employee’s isolated workspace into lasting capabilities for company-owned local models and agents.

Explore development on GitHub