Vehla — Field notes https://vehla.app/Vehla%20Blog.dc.html Comparisons, setup guides, and the reasoning behind the decisions in Vehla. en Sun, 16 Aug 2026 00:00:00 GMT A week of ⌥Space: the eleven commands I actually used https://vehla.app/A%20Week%20of%20Option%20Space.dc.html https://vehla.app/A%20Week%20of%20Option%20Space.dc.html Sun, 16 Aug 2026 00:00:00 GMT Vehla ships 173 commands. A one-week log says eleven of them account for almost everything, half are not AI, and the right move is to hide the rest. Vehla vs Raycast vs SuperCmd: a launcher comparison that admits its own bias https://vehla.app/Vehla%20vs%20Raycast%20vs%20SuperCmd.dc.html https://vehla.app/Vehla%20vs%20Raycast%20vs%20SuperCmd.dc.html Wed, 12 Aug 2026 00:00:00 GMT Three launchers, three business models, three answers to the AI question. The full matrix, the three-year cost of each, and the cases where we would not pick Vehla. QuickGlass, Magic Rewrite, and Flow: three ways to never leave the app https://vehla.app/Three%20Ways%20to%20Never%20Leave.dc.html https://vehla.app/Three%20Ways%20to%20Never%20Leave.dc.html Sat, 08 Aug 2026 00:00:00 GMT Three surfaces that act on text without opening a chat window. Which to use when, plus the permissions each one needs. Vehla vs Alfred: fifteen years of workflows versus a built-in model https://vehla.app/Vehla%20vs%20Alfred.dc.html https://vehla.app/Vehla%20vs%20Alfred.dc.html Tue, 04 Aug 2026 00:00:00 GMT Alfred's workflow ecosystem is the deepest automation layer on macOS and nothing we ship replaces it. What Vehla changes is what happens before you need a workflow. Local AI on your Mac: what actually stays on the device https://vehla.app/Local%20AI%20on%20Your%20Mac.dc.html https://vehla.app/Local%20AI%20on%20Your%20Mac.dc.html Tue, 28 Jul 2026 00:00:00 GMT A precise account of which actions never touch a network, which always do, and why vision is the exception. Recall: what a local index is actually good for https://vehla.app/Recall%20in%20Depth.dc.html https://vehla.app/Recall%20in%20Depth.dc.html Fri, 24 Jul 2026 00:00:00 GMT The queries that work, the four kinds that don't, why citations are the actual feature, and why enabling more sources usually makes answers worse. Six ways to give Vehla a model: catalog, MLX, BYOK, Ollama, LM Studio, Lekh https://vehla.app/Connecting%20AI%20Providers.dc.html https://vehla.app/Connecting%20AI%20Providers.dc.html Sat, 18 Jul 2026 00:00:00 GMT Every route Vehla can take to a language model, what each one needs from you, and which one to pick for a work laptop that has to pass a privacy review. One app instead of fourteen, and the ways that idea usually fails https://vehla.app/Why%20One%20App.dc.html https://vehla.app/Why%20One%20App.dc.html Thu, 09 Jul 2026 00:00:00 GMT Vehla started as a personal tool. Bundling fourteen utilities into one app is the obvious move and also how most bundles get worse than the things they replaced.