🔒
Local execution
Runs on your hardware or a secure internal environment. No internet needed to run, once installed.
Local-first architecture
ChatQP runs on a workstation or a secure internal machine. The model is local. The regulatory corpus is local. The search index is local. Your prompts and batch context do not leave the machine.
What "local" means
Batch records, investigation reports, and quality data stay on your network. There is no cloud processing and no tracking of your queries. For a QP (Qualified Person) or QA (Quality Assurance) team this is the single strongest point in its favour, and it lines up directly with the data control expected in draft Annex 22. First-time setup pulls the local model and corpus onto the machine; after that, running ChatQP needs no internet and your data never leaves it.
Runs on your hardware or a secure internal environment. No internet needed to run, once installed.
The full regulatory corpus is stored and indexed locally. Nothing is fetched from the open web.
Inference and retrieval run on-device. Your inputs are never transmitted off the machine.
Deployment levels
| Level | Purpose |
|---|---|
| Developer workstation | Testing and benchmarking. |
| QA sandbox | Constructed scenarios and training. |
| Validated internal environment | Future controlled deployment. |
| Air-gapped GMP network | Long-term target, subject to validation. |
Depth of analysis
ChatQP offers several depths of analysis, from a fast orientation to a deep multi-domain review distilled into a QP decision.
ChatQP is not supplied as a validated GMP system out of the box. Any regulated deployment would require local validation, SOPs, access control, data governance, change control, and user training.