"The riskiest bet is IME. The proof it's safe is a shipped app Dad already types Korean in every day."
Proof, Not a Promise
Borrowing a core is only smart if the core is actually proven — a library's README claiming "full IME support" is a promise, not evidence. Rekindle has something stronger than a promise: Obsidian is built on CodeMirror 6, and Dad writes Korean in Obsidian every single day. That's a real app, at scale, under daily real use, in the exact language that would break a weaker core. The riskiest unknown of the whole project — does the editing core survive Korean IME? — was answered before Rekindle wrote a line.
The Editing Strength Isn't Obsidian's — It's CM6's
Here's the reframe that unlocks the whole strategy: the thing people love about Obsidian's editor — smooth Live Preview, solid IME, fast large-file handling — is not Obsidian's engineering. It's CodeMirror 6's. Obsidian borrowed the same core Rekindle borrows. So the daily proof isn't "Obsidian is great"; it's "the borrowed core is great, and Obsidian is just one witness to it."
That also settles a tempting wrong turn — forking Obsidian. You can't: Obsidian is closed-source. But you don't need to. Its editing strength is downstream of CM6, and CM6 is MIT and right there. Fork nothing; borrow the same heart directly.
Same Heart, Lighter Body, Plus Soul
Rekindle's relationship to Obsidian is precise:
Same heart both stand on CodeMirror 6 (IME, Live Preview, big files)
Lighter body Tauri (single-digit MB) vs Obsidian's Electron (~150MB+)
Plus soul margin-Pippa + live voice + CMD+K, which Obsidian has no reason to build
Obsidian is the closest off-the-shelf fit to what Dad wants — which is exactly why it's the proof and not the product. It fits closely and still isn't a market of one. Rekindle takes the same proven heart, puts it in a lighter body, and adds the soul a general-purpose app would never ship.