MAIOS / Lab
Research Lab
Autonomous research, system observation and evolution through cycles.
The Lab carries research or monitoring work forward through observation cycles. It can address different questions, follow a system over time and use what emerges to change the work that follows.
The Lab can organise itself as a kernel, bringing context, sources, competences and operating methods together. It keeps the name Lab because its work takes shape through research, observation and the return of results.
Alpha 0.1.0a2 · MIT license · AI in your own environment; Python 3.11+ for local support.
The relations change together: what emerges can transform the context, competences and way of observing.
A reusable Lab in your workspace
Lab 0.1.0a2 provides a semantic core for an AI receiver and local Python support under the MIT license. The AI understands sources and forms the work; the support preserves versions, exact resultants, contacts and selective memory. The package starts with its own state, without the original laboratory’s corpus or private sources.
Enter through the instructions in your AI environment. Local support requires Python 3.11 or later, with no runtime dependencies. A cognitive bridge can connect your chosen model and support bounded autonomous occurrences within your environment’s means and authority.
Questions to explore and systems to follow
A Lab can receive a question to investigate or a system to observe. Research can give rise to further questions; monitoring can reveal a change that calls for a different interpretation. Available sources, tools and conditions participate in how the work takes shape.
The domain can be physics, the operation of an agentic system or another research context. The Lab form remains open to varied questions and can develop methods suited to its work.
Autonomy in the movement of research
The operator can define a context, provide resources and entrust work to the Lab. During its cycles, the Lab forms observations, connects sources and earlier results, and can bring the next question into view. Autonomy concerns this movement of research.
The relationship with the operator continues through monitoring and maintenance: a new source, correction or intervention on the system can change the conditions from which the Lab resumes.
The work also transforms the Lab
A new observation can change what the Lab recognises, the method it uses to address a question or the competences it brings into the next cycle. This capacity to regenerate its operating form is the Lab’s autopoietic relation.
History remains recoverable, while relevant knowledge and capabilities participate in the present. A result can therefore become both research material and a change in how the work continues.
The possibility of operating as an Entity
The direction includes Labs capable of operating as an Entity: forms with their own continuity of observation, understanding and action. How this possibility takes shape remains open to the work and evolution of individual instances.
The name Lab remains even when its organisation develops as a kernel. Each Lab retains its own context and path; different instances can be at different stages of evolution.
A competence can contribute to a different question
In a synthetic exercise with fresh AI sessions, the receiver formed a calibration competence and recovered it from Lab memory for a different monitoring question. Once instrument continuity was explicit, it correctly calculated five new values and their thresholds. Without that competence, on the same source, the result remained indeterminate.
The exercise observes local, guided reuse on the same AI service, with explicit competence selection. When the instrument mode changed, the receiver recognised the calibration’s limit and left the result open. The public repository describes the method, data and scope of the exercise.
An inspectable instance: the D-ND Physics Lab
The D-ND Physics Lab makes a history of occurrences public: finite observations, resultants, relations and recorded contacts with competences. The workspace lets visitors explore these records and follow their provenance; THIA helps examine the selected occurrence and its accompanying relations.
At the 30 October 2026 inspection, the public corpus contains 729 occurrences, with the latest recorded event dated 25 September. The published history makes the process available for study; scientific statements in the records retain their own scope and require the relevant verification.
Explore the corpus and public tools
The dnd-lab-physics repository provides documentation and tools for reading the corpus on a local computer: verify the file, inspect an occurrence, follow recorded contacts and generate an HTML report. These functions require Python 3.11 or later and the public data.
The reusable Lab kernel, the dnd-lab-physics reading tools and the D-ND_LAB engine are distinct distributions. The original laboratory retains its own sources and operational continuity.
Bring Lab into your environment
Download the package and open it in your AI workspace. AGENTS and BOOT lead into the Lab instructions; you can entrust it with a question, context or system to observe. Python support preserves sources, events and competences in a local state directory you choose.
Explore the Physics Lab · Download the public corpus · Read documentation and code
All MAIOS products →