Repository navigation
feat(ai): report local AI hardware and resource compatibility #82
Description
Give users a clear compatibility result before they install or load a local model.
Scope
- Detect supported architecture and runtime capabilities.
- Check available memory and disk space against model requirements.
- Distinguish supported, supported with limits, and unsupported states.
- Keep the check local and collect no telemetry.
Acceptance criteria
- Results include a short reason and an actionable recommendation.
- Boundary values and unavailable system information are tested.
- A failed check never affects the existing PDF tools.
Activity
glte42365-boop commentedon Sep 4, 2026
Hi @McanKul, I would love to work on this issue. I can implement a clean, efficient script to check the available system memory (RAM) and disk space to ensure local model compatibility, while handling supported, limited, and unsupported states without collecting any telemetry. Please assign this issue to me so I can get started!
McanKul commentedon Sep 5, 2026
Thanks @glte42365-boop, assigned to you! Please target
epic/local-ai-runtimefor this AI work. Keep the check in the Rust/Tauri backend and cover architecture/runtime support as well as available RAM and disk space, with model requirements passed in rather than tied to one model. Please test boundary values and unavailable system information, and keep the existing PDF tools working if the check fails. No telemetry or model downloads needed here. Happy to help if anything is unclear!McanKul commentedon Sep 13, 2026
Hi @glte42365-boop, are you still planning to work on this? No pressure—just let us know within the next few days so we can keep the issue available if your plans changed.
feat(ai): report local AI hardware and resource compatibility (#82)
Merge pull request #137 from 1novas/issue#82
McanKul commentedon Oct 22, 2026
PR #137 is merged into
epic/local-ai-runtimeas the local compatibility backend. The import/load UI does not call the check yet, so #82 stays open for the user-visible preflight before a model is loaded or installed. The current Fake runtime correctly reports unsupported until a real runtime is added. Thanks to @1novas for the foundation.