ShellNewHandler VS Tavern.app

Tavern.app's answer:

The wiki fills itself in from the notes your table already takes. Anyone at the table writes a session note; a name or a place in it links to that thing's own wiki page, with backlinks to every session it appeared in. Nobody builds a structure first. Everything else hangs off that loop: a map with pins, character sheets built by hand or synced from the sheet service you already use, private and shared player notebooks, and a Combat Tracker where the GM changes a hit point once and it changes on the sheet the player is already watching. It runs in the browser on a phone with nothing to install, every seat is free with no cap on campaigns or players, and the one optional paid part, Lorekeeper, writes the wiki entries and map pins from the GM's own notes and is off unless the GM sends it a session.

Tavern.app's answer:

Because the whole table can be on it without anyone paying and without anyone doing homework. Most campaign wikis cap a free campaign at eight to twelve members or sell player slots by the owner's tier; a Tavern campaign has no seat cap and players never pay. Most of them also ask you to create entities and folders before they are useful; Tavern asks you to write a session note and makes the pages from it. It reads well on a phone under the table, which is where sessions happen. And it takes a clear stance on trust: it ships none of any publisher's content, never trains AI on what a table writes, and keeps the one AI feature opt-in and off by default. If you want your notes as files on your own disk, use Obsidian. If you want the table inside the notes, this is the one.

Tavern.app's answer:

Game Masters running an ongoing tabletop roleplaying campaign, any system, and the four or five players at their table. The GM is the person who starts a campaign and hands out the link; the players are the people who open it on a phone on the way over. It suits long campaigns most: a table that meets every week or two, keeps forgetting who that merchant was, and has tried a shared doc or a wiki app and drifted off it. College clubs and other groups that start several campaigns at once are a natural fit too, since a new table gets one link and a shared place to write from session one.

Tavern.app's answer:

I run a table in Nashville and I was the one person at it who remembered anything. Every session opened with me retelling the last one to people who had been there. I tried a shared doc, a wiki app and a notes app with plugins; each one asked the table to build a structure before it paid off, and each one lived on a laptop while my players held phones. So I built the thing I wanted: a session journal the whole table writes in, where a name in a note becomes a page, on a phone, free for every seat. My own campaign ran on it first. Sign-ups opened in August 2026, and I want to build the rest of it alongside the people using it: tell me what's awesome, what sucks and what's missing.

Tavern.app's answer:

TypeScript throughout. React 19 with Vite and Tailwind, framer-motion for motion and TanStack Query for data, served as a PWA that installs per campaign. A Hono API on Cloudflare Workers, with D1 (SQLite) for the journals and R2 for uploaded pictures, all on the Workers free plan. Share cards and the icon set's PNGs are rasterised in the Worker with satori and resvg. Mail goes through Resend. Lorekeeper, the optional AI feature, is a scheduled Claude agent that claims jobs from a queue in D1 and lands each session as a versioned database migration, so every wiki page it touches keeps a revision snapshot. Tests in vitest, including one that fails the build if any publisher's content or a game's name ever gets into the product.

Tavern.app's answer:

• My own table in Nashville, the first campaign in the database
• A handful of home tables and college groups who signed up in the first week; they are private groups and I don't name them