What you can inspect
- The source excerpts behind an answer
- The selected processing mode and provider path
- The distinction between canonical and derived data
Search and reason over portable knowledge with local-first retrieval, inspectable sources, and an explicit choice of local, connected, or hosted model paths.
Your provider remains the canonical home for Markdown. Derived indexes can be rebuilt, and connected models receive bounded selected context only when you choose that route.
Phasoric separates your canonical files, rebuildable retrieval data, and optional model processing so the privacy and provider tradeoff is understandable before a request runs.
Markdown stays readable in the storage provider you select; an intelligence index does not become a replacement source of truth.
Supported search and intelligence paths begin on the device and keep the matching source context visible.
Use supported local processing or deliberately send bounded selected context through a connected or hosted model route.
Model availability and speed vary by device, browser, plan, and provider. Phasoric keeps that limitation visible at the point of use.
No. Supported retrieval and some intelligence can run locally, while connected or hosted models are optional processing paths. When a connected model is selected, bounded relevant context is sent to that provider.
Yes. Phasoric supports local and mapped provider paths, including supported hosted, Dropbox, Google Drive, and GitHub workflows. Availability varies by capability and plan.
Phasoric keeps source context and citations available so you can return to the original material. Generated output remains a draft or suggestion until a person reviews it.
Keep portable knowledge at the center while you decide which intelligence path is appropriate for each task.
Start with local intelligence