My Emacs configuration https://git.cjennings.net/dotemacs/atom?h=load-graph-classify-start 2026-05-14T06:26:28+00:00 2026-05-14T06:26:28+00:00 Craig Jennings c@cjennings.net 2026-05-14T06:26:28+00:00 urn:sha1:a9a9e56148e9ab83f9abc4e4c531be86974364d1 Second pass on ai-config. First batch covered the auth-source, API-key caching, file-context, clear-buffer, context-clear, and model-heading helpers. This batch covers the gptel command wrappers and a couple of small helpers: - `cj/gptel--refresh-org-prefix`: refreshes the org-mode entry in `gptel-prompt-prefix-alist`. - `cj/gptel-backend-and-model`: formats a vector-backend's name + the current model + a timestamp; falls back to "AI:" when no backend is bound. - `cj/gptel-switch-backend`: picks a backend and one of its models, updates both gptel vars; errors on an invalid backend choice. - `cj/gptel-add-buffer-file`: hands the visited file off to `cj/gptel--add-file-to-context`, messages otherwise. - `cj/gptel-add-this-buffer`: calls `gptel-add` with the (4) prefix-arg. - `cj/toggle-gptel`: deletes the window when *AI-Assistant* is already shown. gptel and projectile primitives are stubbed throughout.