Under testing · Legacy to AI

SLM Centauri™

Helping organisations move existing data towards an AI-ready future. Centauri is being developed to bring data transfer and a connected, evidence-backed knowledge layer together.

From legacy information to AI-ready knowledge

Organisations hold valuable information across files, folders, systems and shared storage. SLM Centauri is intended to help bring that information together while retaining the context and evidence needed to understand where it came from.

The goal is a connected, portable knowledge layer that authorised people and AI tools can use to explore information and relationships, with source evidence available for review. Centauri is under testing; not every planned capability has been independently qualified.

The intended workflow

1. Choose the source

Select an authorised folder, drive, server share or other supported source.

2. Choose the destination

Choose where the transferred information and its knowledge records should be kept.

3. Start the operation

Centauri is intended to transfer source data while collecting supported content, metadata and evidence. The design aims to leave the original source unchanged.

4. Build connected knowledge

Supported information can be organised into facts and relationships with source references, so users can trace a result back to the underlying evidence.

5. Review and use

Authorised users and tools should be able to explore the resulting knowledge. Permission-aware access, supported formats and integration behaviour remain part of validation.

Design aim: account for discovered items and record exceptions, including unsupported, inaccessible or failed items, rather than quietly presenting an incomplete result as complete.

What Centauri is being built to support

Source and provenance

Capture source context and evidence so that supported findings can be traced and reviewed.

Connected information

Organise supported facts and relationships across documents, systems and other authorised sources.

Customer-controlled environments

The intended core workflow uses the organisation's own infrastructure and is designed to work without a required cloud service or LLM.

Permissions and boundaries

The design calls for permission-aware access and separation of different estates. Full enforcement remains subject to implementation and validation.

Recovery and outcomes

The intended workflow records item outcomes and uses checkpoints to support recovery from interrupted work.

What has been tested

The following are bounded development results from recorded Centauri/LegacyAI test artefacts. They describe individual checks and are not a claim that the whole product is production-ready.

Legacy Office document processing

Recorded checks passed for selected DOC, XLS and PPT fixtures, as well as embedded-object, adversarial-format and helper-failure isolation cases. A package smoke check also passed for three sample legacy Office files.

Knowledge queries and recovery

A focused later run passed checks for large-query page transport, history pagination and rollback, and pagination with provenance. An earlier run had recorded a rollback failure; the later focused result is retained separately.

Package integrity

A recorded archive verification reported a valid manifest covering 497 components. That result applies to the particular tested archive.

Estate and scale work

Test artefacts include a 2 GB estate run and a recorded job-start response. A completed end-to-end result for that run has not been established from the records reviewed for this page.

Further validation is needed across supported formats, permissions, real environments, recovery and deployment. Centauri remains under testing; these results should not be treated as a blanket guarantee for every data source or environment.

Interested in Centauri?

Centauri is under testing. Contact SLM-AI Solutions to discuss the project or testing opportunities.

Contact us →