yt-briefing 1.0.0 → 1.1.1

package/README.md

@@ -61,8 +61,9 @@ YT_BRIEFING_YOUTUBE_API_KEY=<key>    # console.cloud.google.com → enable "YouT
 ```
 
 Optional extras: `YT_BRIEFING_MODEL` picks the Claude model for filtering and summaries (default
-`haiku`, any alias or model name `claude --model` accepts), and `YT_BRIEFING_PROXY` routes
-transcript fetches through a proxy on datacenter/VPS IPs.
+`sonnet`, any alias or model name `claude --model` accepts, `haiku` to go lighter on your plan's
+usage limits), and `YT_BRIEFING_PROXY` routes transcript fetches through a proxy on
+datacenter/VPS IPs.
 
 4. Onboard:
 
@@ -122,6 +123,9 @@ it to go deeper, lower it for a quicker pass:
 Open your project in Claude Code and type `/yt`. The briefing opens in a pane: the summary, and
 under it four keys.
 
+The keys work while the pane has the focus (it takes it when it opens, `/yt` from an empty
+prompt), and Esc closes it.
+
 | Key | What it does |
 |-----|--------------|
 | `1` OK | Neutral. The video is marked as seen, the next one loads. |
@@ -132,9 +136,12 @@ under it four keys.
 The **Comment** field takes anything else. Type what you think in your own words ("too many panel
 shows, skip those") and press Enter: Claude turns it into a standing rule for that channel and
 infers the rating. `? your question` starts research with that question, `stop` closes the pane.
+Claude Code on a phone has no text fields, so there the pane shows the four keys only.
 
-Each step is the engine, not a chat turn: rating a video costs no tokens of your session, and
-the next summary is usually ready before you have finished reading the current one.
+Each step is the engine, not a chat turn: rating a video takes no turn and no context of your
+session, and the next summary is usually ready before you have finished reading the current one.
+The summaries themselves still count toward your Claude plan's usage, like any other Claude Code
+work.
 
 `/yt` is a Claude Code mod (a plugin in `.claude/skills/yt-briefing/`). Claude Code loads it on
 its own once you trust the project folder. If `/yt` is not listed, start a fresh session. To
@@ -195,6 +202,11 @@ at a separate private repo) and commit after each rating: set `"after_rate"` in
 
 ## Running on a VPS
 
+Claude Code has to be installed and logged in on the server too, because the engine runs
+`claude -p` there. With no browser on the box, `claude setup-token` creates a long-lived login
+token for your subscription. An `ANTHROPIC_API_KEY` alone is not enough: the engine removes it
+for its calls (see above), so a server set up only with an API key fails with a login error.
+
 YouTube blocks datacenter IPs, so transcript fetches fail on most servers. Route them through a
 free Cloudflare WARP proxy. See [docs/warp-proxy.md](./docs/warp-proxy.md).
 

package/data.example/README.md

@@ -5,7 +5,8 @@ reference for the shape of that data, and a home for the channel-profile templat
 
 ```
 data/
-├── config.json          { "output_lang": "English" }   ← your summary/rating language
+├── config.json          { "output_lang": "English" }   ← your summary language. Optional
+│                        "after_rate": "<command>" runs after every rating (sync)
 ├── channels.md          the channels you follow (a flat list)
 ├── state.md             per-channel per-type cursor (last video seen of each type)
 ├── channels/

package/dist/lib/llm.js

@@ -15,11 +15,15 @@
  * (or a "credit balance too low" failure). The briefing is meant to run on the login.
  *
  * Env (all optional):
- *   YT_BRIEFING_MODEL   model alias or full name passed to `claude --model` (default: haiku)
+ *   YT_BRIEFING_MODEL   model alias or full name passed to `claude --model` (default: sonnet)
  */
 import { spawn, spawnSync } from 'node:child_process';
-/** Model alias for every call; one model does both stages, as before. */
-export const DEFAULT_MODEL = 'haiku';
+/**
+ * Model alias for every call; one model does both stages. Sonnet: the summary is the product, and a
+ * subscription login does not bill per token, so the default buys quality over speed. Set
+ * YT_BRIEFING_MODEL=haiku for faster, lighter runs.
+ */
+export const DEFAULT_MODEL = 'sonnet';
 export function getModel() {
     return process.env.YT_BRIEFING_MODEL || DEFAULT_MODEL;
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "yt-briefing",
-  "version": "1.0.0",
+  "version": "1.1.1",
   "description": "A self-learning YouTube briefing for Claude Code: it sweeps the channels you follow, filters noise in two stages (title, then transcript), gives you a short briefing of the rest in your language in a pane, and adapts to your ratings — one video at a time.",
   "type": "module",
   "bin": {

package/plugin/.claude-plugin/plugin.json

@@ -1,7 +1,9 @@
 {
   "name": "yt-briefing",
-  "version": "1.0.0",
+  "version": "1.1.1",
   "description": "The /yt briefing loop as a Claude Code pane: summary, one-key rating, research hand-off.",
-  "author": { "name": "Michal Ryzio" },
+  "author": {
+    "name": "Michal Ryzio"
+  },
   "types": "./types/index.d.ts"
 }