Aren’t inline predictive text and Apple Intelligence the same thing?
No, they’re two separate features. Inline predictive text is the short grey suggestion that arrived in macOS Sonoma — it finishes the current word or a few words and mostly works in Apple’s own apps (Notes, Mail, Messages, Pages). Apple Intelligence is the newer Writing Tools layer (Proofread, Rewrite, tone, Summarize) that transforms text you select. Neither gives you continuous iOS-keyboard-style ghost prediction in every text field — that’s what Shadowtype adds.
If Apple Intelligence is free, why pay for Shadowtype?
Because they do different jobs. Apple Intelligence is excellent for rewriting selected text for free. Shadowtype is a focused tool for the thing Apple doesn’t do: continuous inline prediction in any app, guaranteed 100% on-device, with model choice and per-app tuning. Many people run both.
Is Apple Intelligence really not fully on-device?
Apple runs many requests on-device, but Writing Tools can route larger tasks to Private Cloud Compute — Apple’s own servers — and the optional ChatGPT extension goes off-device entirely. Shadowtype guarantees zero network for completion, so nothing leaves your Mac even if you’re offline.
Do they run on the same Macs?
Both require Apple Silicon. The difference is the OS: Apple Intelligence needs macOS 15.1 Sequoia or later, while Shadowtype runs on macOS 14 — so it works on supported Macs a full OS version earlier.
Can my own scripts, editors, and agents use Shadowtype's local model?
Yes. Shadowtype exposes an OpenAI-compatible HTTP endpoint on 127.0.0.1 with no key required, so your terminal, editor plugins, and AI agents can call /v1/chat/completions against the same on-device model. A built-in MCP server plugs into Claude Code and other MCP clients, and the BYOM picker loads any GGUF you drop in. Apple Intelligence keeps its on-device models locked to Apple’s own apps and Writing Tools — there’s no localhost API or MCP server for your code.