Subagents: one slider with two handles, Min and Max; 1 is the agent alone

The chip's popover is one range from 1 to the server's cap (subagents.cap, /models
subagents_cap, else 32) with two handles: two range inputs overlaid on one track, the span
filled, aria-labels Min and Max with value text, keyboard as any slider, Min never past Max,
the values said above it ("Min 1 · Max 4") and the two ends under it. "Fewest"/"Most" are
gone. The hint says what the numbers count: the chat's own agent -- 1 is the agent alone, Max 4
is the agent and up to 3 subagents. Untouched on a new chat it still leaves the server's
defaults (1-4, shown). i18n: Min, Max and the new hint in all eight languages. CHANGELOG.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

CHANGELOG.md

@@ -9,7 +9,10 @@ All notable changes to LLeMbas Desktop. The format follows
 ### Changed
 
 - **Subagents**, by that name everywhere. The chip under the message box says **Subagents** with the
-  fewest and the most a chat runs, and opens the two fields under those words; it shows where the
+  least and the most agents a chat runs ("Subagents 1–4"), and opens one slider with two handles,
+  **Min** and **Max**, from 1 to the server's cap -- Min never passes Max, and both move with the
+  arrow keys as well. The numbers count the chat's own agent: **1 is the agent alone**, with no
+  subagents; Max 4 is the agent and up to 3 subagents. The chip shows where the
   server offers subagents for the chat (for a new chat, for its model), and the most is the
   server's cap. Where the chat's model serves one request at a time, it says that subagents need
   another model, on another server, named in the chat. Subagents work on the chat's own model

scripts/gen-i18n.py

@@ -194,8 +194,8 @@ KEYS = (
     "off",
     "default",
     "Think before answering",
-    "Fewest",
-    "Most",
+    "Min",
+    "Max",
     "Continue in a terminal…",
     "Continue in a terminal",
     "Run this in a terminal: the agent's session goes on there, in this chat's folder, and the chat keeps it.",
@@ -367,7 +367,7 @@ KEYS = (
     "Subagent",
     "Team member",
     "Subagents",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.",
     "This model serves one request at a time: subagents need another model, on another server, named in the chat.",
     "None",
     "Up to %(n)s at once",
@@ -382,7 +382,9 @@ KEYS = (
 OWN: dict[str, dict[str, str]] = {
     "cs": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Tento model vyřizuje najednou jeden požadavek: podagenti potřebují jiný model na jiném serveru, pojmenovaný v konverzaci.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Počítá se včetně vlastního agenta konverzace. Podagent běží na modelu konverzace, na jiném jen tehdy, když ho v konverzaci pojmenujete.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Čísla počítají i vlastního agenta konverzace: 1 je agent sám, bez podagentů; Max 4 je agent a až 3 podagenti. Podagent běží na modelu konverzace, na jiném jen tehdy, když ho v konverzaci pojmenujete.",
+        "Min": "Min",
+        "Max": "Max",
         "Tasks: %(n)s": "Úlohy: %(n)s",
         "%(n)s running": "Běží: %(n)s",
         "%(tokens)s until auto-compact": "%(tokens)s do automatického shrnutí",
@@ -601,7 +603,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "de": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Dieses Modell bedient eine Anfrage zur Zeit: Subagenten brauchen ein anderes Modell auf einem anderen Server, im Chat genannt.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Der eigene Agent des Chats zählt mit. Ein Subagent läuft auf dem Modell des Chats, auf einem anderen nur, wenn Sie es im Chat nennen.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Die Zahlen zählen den eigenen Agenten des Chats mit: 1 ist der Agent allein, ohne Subagenten; Max 4 ist der Agent und bis zu 3 Subagenten. Ein Subagent läuft auf dem Modell des Chats, auf einem anderen nur, wenn Sie es im Chat nennen.",
+        "Min": "Min",
+        "Max": "Max",
         "Tasks: %(n)s": "Aufgaben: %(n)s",
         "%(n)s running": "%(n)s laufen",
         "%(tokens)s until auto-compact": "%(tokens)s bis zur automatischen Zusammenfassung",
@@ -820,7 +824,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "es": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Este modelo atiende una petición a la vez: los subagentes necesitan otro modelo, en otro servidor, nombrado en el chat.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Cuenta también el agente propio del chat. Un subagente usa el modelo del chat, u otro solo si lo nombra en el chat.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Los números cuentan el agente propio del chat: 1 es el agente solo, sin subagentes; Máx. 4 es el agente y hasta 3 subagentes. Un subagente usa el modelo del chat, u otro solo si lo nombra en el chat.",
+        "Min": "Mín.",
+        "Max": "Máx.",
         "Tasks: %(n)s": "Tareas: %(n)s",
         "%(n)s running": "%(n)s en marcha",
         "%(tokens)s until auto-compact": "%(tokens)s hasta la compactación automática",
@@ -1039,7 +1045,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "fr": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Ce modèle traite une requête à la fois : les sous-agents ont besoin d’un autre modèle, sur un autre serveur, nommé dans la conversation.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "L’agent de la conversation compte aussi. Un sous-agent utilise le modèle de la conversation, ou un autre seulement si vous le nommez dans la conversation.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Les nombres comptent l’agent de la conversation : 1, c’est l’agent seul, sans sous-agents ; Max 4, c’est l’agent et jusqu’à 3 sous-agents. Un sous-agent utilise le modèle de la conversation, ou un autre seulement si vous le nommez dans la conversation.",
+        "Min": "Min",
+        "Max": "Max",
         "Tasks: %(n)s": "Tâches : %(n)s",
         "%(n)s running": "%(n)s en cours",
         "%(tokens)s until auto-compact": "%(tokens)s avant le compactage automatique",
@@ -1258,7 +1266,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "hu": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Ez a modell egyszerre egy kérést szolgál ki: az alügynököknek másik modell kell, másik szerveren, a beszélgetésben megnevezve.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "A beszélgetés saját ügynöke is beleszámít. Az alügynök a beszélgetés modelljén fut, másikon csak akkor, ha megnevezi a beszélgetésben.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "A számokba a beszélgetés saját ügynöke is beleszámít: az 1 maga az ügynök, alügynökök nélkül; a Max. 4 az ügynök és legfeljebb 3 alügynök. Az alügynök a beszélgetés modelljén fut, másikon csak akkor, ha megnevezi a beszélgetésben.",
+        "Min": "Min.",
+        "Max": "Max.",
         "Tasks: %(n)s": "Feladatok: %(n)s",
         "%(n)s running": "%(n)s fut",
         "%(tokens)s until auto-compact": "%(tokens)s az automatikus tömörítésig",
@@ -1477,7 +1487,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "it": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Questo modello serve una richiesta alla volta: i sottoagenti hanno bisogno di un altro modello, su un altro server, nominato nella chat.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Conta anche l’agente della chat. Un sottoagente usa il modello della chat, o un altro solo se lo nomini nella chat.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "I numeri contano anche l’agente della chat: 1 è l’agente da solo, senza sottoagenti; Max 4 è l’agente più fino a 3 sottoagenti. Un sottoagente usa il modello della chat, o un altro solo se lo nomini nella chat.",
+        "Min": "Min",
+        "Max": "Max",
         "Tasks: %(n)s": "Attività: %(n)s",
         "%(n)s running": "%(n)s in corso",
         "%(tokens)s until auto-compact": "%(tokens)s alla compattazione automatica",
@@ -1696,7 +1708,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "pl": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Ten model obsługuje jedno żądanie naraz: podagenci potrzebują innego modelu, na innym serwerze, wskazanego w czacie.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Liczony razem z własnym agentem czatu. Podagent działa na modelu czatu, na innym tylko wtedy, gdy wskażesz go w czacie.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Liczby obejmują własnego agenta czatu: 1 to sam agent, bez podagentów; Maks 4 to agent i do 3 podagentów. Podagent działa na modelu czatu, na innym tylko wtedy, gdy wskażesz go w czacie.",
+        "Min": "Min",
+        "Max": "Maks",
         "Tasks: %(n)s": "Zadania: %(n)s",
         "%(n)s running": "Działa: %(n)s",
         "%(tokens)s until auto-compact": "%(tokens)s do automatycznego streszczenia",
@@ -1915,7 +1929,9 @@ OWN: dict[str, dict[str, str]] = {
     },
     "sk": {
         "This model serves one request at a time: subagents need another model, on another server, named in the chat.": "Tento model vybavuje naraz jednu požiadavku: podagenti potrebujú iný model na inom serveri, pomenovaný v konverzácii.",
-        "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Počíta sa vrátane vlastného agenta konverzácie. Podagent beží na modeli konverzácie, na inom len vtedy, keď ho v konverzácii pomenujete.",
+        "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Čísla počítajú aj vlastného agenta konverzácie: 1 je agent sám, bez podagentov; Max 4 je agent a až 3 podagenti. Podagent beží na modeli konverzácie, na inom len vtedy, keď ho v konverzácii pomenujete.",
+        "Min": "Min",
+        "Max": "Max",
         "Tasks: %(n)s": "Úlohy: %(n)s",
         "%(n)s running": "Beží: %(n)s",
         "%(tokens)s until auto-compact": "%(tokens)s do automatického zhrnutia",

ui/app.css

@@ -709,28 +709,75 @@ details[open] > summary .step__chev .icon { transform: rotate(90deg); }
 }
 .toggle input:checked { background: var(--accent); border-color: var(--accent); }
 .toggle input:checked::after { background: var(--accent-ink); transform: translateX(var(--check-size)); }
-/* The subagents, fewest and most: two number fields of one width with a dash between them, and in the header
-   its word before them -- fixed tracks, so nothing moves when a number grows a digit. */
-.agents, .agents__heads {
-  display: inline-grid;
-  grid-template-columns: minmax(0, 1fr) var(--sp-3) minmax(0, 1fr);
-  gap: var(--sp-2);
-  align-items: center;
+/* The subagents, Min and Max: one range with two handles. Three rows of known height -- the
+   values said on one line, the track box (one control high), the two ends -- and in the box two
+   range inputs laid over one track, the span between the handles filled. The inputs let only
+   their handles take the pointer, so either can be grabbed wherever the other is. */
+.range2 { display: grid; grid-template-rows: auto var(--control-h) auto; gap: var(--sp-1); width: 100%; }
+.range2__said { font: 600 var(--text-sm) / var(--leading-tight) var(--font-ui); color: var(--ink); font-variant-numeric: tabular-nums; }
+.range2__box { position: relative; display: block; height: var(--control-h); --lo: 0%; --hi: 100%; }
+.range2__track {
+  position: absolute;
+  top: 50%;
+  left: calc(var(--check-size) / 2);
+  right: calc(var(--check-size) / 2);
+  height: var(--sp-1);
+  transform: translateY(-50%);
+  background: var(--border);
+  border-radius: var(--radius-full);
+}
+.range2__fill {
+  position: absolute;
+  top: 0;
+  bottom: 0;
+  left: var(--lo);
+  right: calc(100% - var(--hi));
+  background: var(--accent);
+  border-radius: var(--radius-full);
+}
+.range2__input {
+  position: absolute;
+  inset: 0;
   width: 100%;
-  height: var(--control-h);
+  height: 100%;
+  margin: 0;
+  background: none;
+  pointer-events: none;
+  -webkit-appearance: none;
+  appearance: none;
 }
-.agents__dash { color: var(--ink-muted); text-align: center; }
-/* The words over the subagents' two fields, on the fields' own tracks. */
-.agents__heads { height: auto; font-size: var(--text-xs); color: var(--ink-muted); }
-.agents__input { width: 100%; min-width: 0; height: var(--control-h); padding: 0 var(--control-px-sm); }
-.agents--compact {
-  grid-template-columns: auto calc(var(--w-field-xs) / 2) auto calc(var(--w-field-xs) / 2);
-  width: auto;
-  height: var(--control-h-sm);
-  font: var(--text-sm) var(--font-ui);
+.range2__input.is-top { z-index: 1; }
+.range2__input::-webkit-slider-runnable-track { height: 100%; background: none; border: 0; }
+.range2__input::-moz-range-track { height: 100%; background: none; border: 0; }
+.range2__input::-webkit-slider-thumb {
+  -webkit-appearance: none;
+  appearance: none;
+  width: var(--check-size);
+  height: var(--check-size);
+  margin-top: calc((var(--control-h) - var(--check-size)) / 2);
+  background: var(--surface-raised);
+  border: var(--border-w-thick) solid var(--accent);
+  border-radius: var(--radius-full);
+  box-shadow: var(--shadow-sm);
+  cursor: grab;
+  pointer-events: auto;
 }
-.agents--compact .agents__input { height: var(--control-h-sm); }
-.agents__label { color: var(--ink-muted); white-space: nowrap; }
+.range2__input::-moz-range-thumb {
+  width: var(--check-size);
+  height: var(--check-size);
+  background: var(--surface-raised);
+  border: var(--border-w-thick) solid var(--accent);
+  border-radius: var(--radius-full);
+  box-shadow: var(--shadow-sm);
+  cursor: grab;
+  pointer-events: auto;
+}
+.range2__input:focus-visible { outline: none; }
+.range2__input:focus-visible::-webkit-slider-thumb { box-shadow: var(--ring); }
+.range2__input:focus-visible::-moz-range-thumb { box-shadow: var(--ring); }
+/* The two ends, each one handle wide, so each number sits centred under its handle at that end. */
+.range2__ends { display: flex; justify-content: space-between; font-size: var(--text-xs); color: var(--ink-muted); font-variant-numeric: tabular-nums; }
+.range2__ends > span { width: var(--check-size); text-align: center; white-space: nowrap; }
 /* Markdown, drawn by markdown.js from model output (never as HTML). */
 .md { overflow-wrap: anywhere; min-width: 0; }
 .md > * + * { margin-top: var(--sp-3); }

ui/i18n.js

@@ -83,7 +83,6 @@ const LEMBAS_I18N = {
     "Copied": "Zkopírováno",
     "Copy": "Kopírovat",
     "Counted by the model at the last reply; the categories are estimated.": "Spočítal model při poslední odpovědi; kategorie jsou odhadnuté.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Počítá se včetně vlastního agenta konverzace. Podagent běží na modelu konverzace, na jiném jen tehdy, když ho v konverzaci pojmenujete.",
     "Cut short here; the whole of it is in the web interface.": "Zde zkráceno; celé je to ve webovém rozhraní.",
     "Days": "Dny",
     "Delete": "Smazat",
@@ -109,7 +108,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Každý výsledek, i když není co hlásit.",
     "Every weekday at 8:00, summarise the news on my reading list": "Každý pracovní den v 8:00 shrň zprávy z mého seznamu ke čtení",
     "Failed": "Selhalo",
-    "Fewest": "Nejméně",
     "Files": "Soubory",
     "Finished %(n)s": "Dokončeno: %(n)s",
     "Folder": "Složka",
@@ -134,11 +132,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Hledám tam server LLeMbas…",
     "MCP tools": "Nástroje MCP",
     "Manual — asks first": "Manual — nejdřív se zeptá",
+    "Max": "Max",
     "Maximise": "Maximalizovat",
     "Memory and personal": "Paměť a osobní",
     "Menu": "Nabídka",
     "Message": "Zpráva",
     "Messages": "Zprávy",
+    "Min": "Min",
     "Minimise": "Minimalizovat",
     "Missed": "Zmeškáno",
     "Mode": "Režim",
@@ -146,7 +146,6 @@ const LEMBAS_I18N = {
     "Models": "Modely",
     "Models for scheduled tasks": "Modely pro naplánované úlohy",
     "More": "Více",
-    "Most": "Nejvíce",
     "Never": "Nikdy",
     "New agent chat": "Nová agentní konverzace",
     "New chat": "Nová konverzace",
@@ -269,6 +268,7 @@ const LEMBAS_I18N = {
     "The model's own": "Podle modelu",
     "The most an agent chat may do here": "Nejvíc, co tu agentní konverzace smí",
     "The next runs": "Další spuštění",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Čísla počítají i vlastního agenta konverzace: 1 je agent sám, bez podagentů; Max 4 je agent a až 3 podagenti. Podagent běží na modelu konverzace, na jiném jen tehdy, když ho v konverzaci pojmenujete.",
     "The reply is ready.": "Odpověď je hotová.",
     "The result is no longer there.": "Výsledek už tu není.",
     "The server refused this computer's link: %(why)s": "Server odmítl propojení tohoto počítače: %(why)s",
@@ -412,7 +412,6 @@ const LEMBAS_I18N = {
     "Copied": "Kopiert",
     "Copy": "Kopieren",
     "Counted by the model at the last reply; the categories are estimated.": "Vom Modell bei der letzten Antwort gezählt; die Kategorien sind geschätzt.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Der eigene Agent des Chats zählt mit. Ein Subagent läuft auf dem Modell des Chats, auf einem anderen nur, wenn Sie es im Chat nennen.",
     "Cut short here; the whole of it is in the web interface.": "Hier gekürzt; vollständig steht es in der Weboberfläche.",
     "Days": "Tage",
     "Delete": "Löschen",
@@ -438,7 +437,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Jedes Ergebnis, auch ohne Neuigkeiten.",
     "Every weekday at 8:00, summarise the news on my reading list": "Jeden Werktag um 8:00 die Nachrichten auf meiner Leseliste zusammenfassen",
     "Failed": "Fehlgeschlagen",
-    "Fewest": "Mindestens",
     "Files": "Dateien",
     "Finished %(n)s": "Fertig: %(n)s",
     "Folder": "Ordner",
@@ -463,11 +461,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Suche dort nach einem LLeMbas-Server…",
     "MCP tools": "MCP-Werkzeuge",
     "Manual — asks first": "Manual — fragt zuerst",
+    "Max": "Max",
     "Maximise": "Maximieren",
     "Memory and personal": "Gedächtnis und Persönliches",
     "Menu": "Menü",
     "Message": "Nachricht",
     "Messages": "Nachrichten",
+    "Min": "Min",
     "Minimise": "Minimieren",
     "Missed": "Verpasst",
     "Mode": "Modus",
@@ -475,7 +475,6 @@ const LEMBAS_I18N = {
     "Models": "Modelle",
     "Models for scheduled tasks": "Modelle für geplante Aufgaben",
     "More": "Mehr",
-    "Most": "Höchstens",
     "Never": "Nie",
     "New agent chat": "Neuer Agenten-Chat",
     "New chat": "Neuer Chat",
@@ -598,6 +597,7 @@ const LEMBAS_I18N = {
     "The model's own": "Die des Modells",
     "The most an agent chat may do here": "Das Meiste, was ein Agent-Chat hier darf",
     "The next runs": "Die nächsten Läufe",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Die Zahlen zählen den eigenen Agenten des Chats mit: 1 ist der Agent allein, ohne Subagenten; Max 4 ist der Agent und bis zu 3 Subagenten. Ein Subagent läuft auf dem Modell des Chats, auf einem anderen nur, wenn Sie es im Chat nennen.",
     "The reply is ready.": "Die Antwort ist fertig.",
     "The result is no longer there.": "Das Ergebnis ist nicht mehr da.",
     "The server refused this computer's link: %(why)s": "Der Server hat die Verknüpfung dieses Computers abgelehnt: %(why)s",
@@ -741,7 +741,6 @@ const LEMBAS_I18N = {
     "Copied": "Copiado",
     "Copy": "Copiar",
     "Counted by the model at the last reply; the categories are estimated.": "Contado por el modelo en la última respuesta; las categorías son estimaciones.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Cuenta también el agente propio del chat. Un subagente usa el modelo del chat, u otro solo si lo nombra en el chat.",
     "Cut short here; the whole of it is in the web interface.": "Recortado aquí; completo está en la interfaz web.",
     "Days": "Días",
     "Delete": "Eliminar",
@@ -767,7 +766,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Cada resultado, incluso sin nada que informar.",
     "Every weekday at 8:00, summarise the news on my reading list": "Cada día laborable a las 8:00, resume las noticias de mi lista de lectura",
     "Failed": "Fallidos",
-    "Fewest": "Mínimo",
     "Files": "Archivos",
     "Finished %(n)s": "Terminadas: %(n)s",
     "Folder": "Carpeta",
@@ -792,11 +790,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Buscando un servidor LLeMbas allí…",
     "MCP tools": "Herramientas MCP",
     "Manual — asks first": "Manual — pregunta primero",
+    "Max": "Máx.",
     "Maximise": "Maximizar",
     "Memory and personal": "Memoria y datos personales",
     "Menu": "Menú",
     "Message": "Mensaje",
     "Messages": "Mensajes",
+    "Min": "Mín.",
     "Minimise": "Minimizar",
     "Missed": "Perdida",
     "Mode": "Modo",
@@ -804,7 +804,6 @@ const LEMBAS_I18N = {
     "Models": "Modelos",
     "Models for scheduled tasks": "Modelos para tareas programadas",
     "More": "Más",
-    "Most": "Máximo",
     "Never": "Nunca",
     "New agent chat": "Nuevo chat de agente",
     "New chat": "Nuevo chat",
@@ -927,6 +926,7 @@ const LEMBAS_I18N = {
     "The model's own": "La del modelo",
     "The most an agent chat may do here": "Lo máximo que un chat de agente puede hacer aquí",
     "The next runs": "Las próximas ejecuciones",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Los números cuentan el agente propio del chat: 1 es el agente solo, sin subagentes; Máx. 4 es el agente y hasta 3 subagentes. Un subagente usa el modelo del chat, u otro solo si lo nombra en el chat.",
     "The reply is ready.": "La respuesta está lista.",
     "The result is no longer there.": "El resultado ya no está.",
     "The server refused this computer's link: %(why)s": "El servidor rechazó la vinculación de este equipo: %(why)s",
@@ -1070,7 +1070,6 @@ const LEMBAS_I18N = {
     "Copied": "Copié",
     "Copy": "Copier",
     "Counted by the model at the last reply; the categories are estimated.": "Compté par le modèle lors de la dernière réponse ; les catégories sont estimées.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "L’agent de la conversation compte aussi. Un sous-agent utilise le modèle de la conversation, ou un autre seulement si vous le nommez dans la conversation.",
     "Cut short here; the whole of it is in the web interface.": "Tronqué ici ; le tout est dans l'interface web.",
     "Days": "Jours",
     "Delete": "Supprimer",
@@ -1096,7 +1095,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Chaque résultat, même sans rien à signaler.",
     "Every weekday at 8:00, summarise the news on my reading list": "Chaque jour ouvré à 8 h, résume les nouvelles de ma liste de lecture",
     "Failed": "Échecs",
-    "Fewest": "Au moins",
     "Files": "Fichiers",
     "Finished %(n)s": "Terminées : %(n)s",
     "Folder": "Dossier",
@@ -1121,11 +1119,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Recherche d'un serveur LLeMbas à cette adresse…",
     "MCP tools": "Outils MCP",
     "Manual — asks first": "Manual — demande d'abord",
+    "Max": "Max",
     "Maximise": "Agrandir",
     "Memory and personal": "Mémoire et informations personnelles",
     "Menu": "Menu",
     "Message": "Message",
     "Messages": "Messages",
+    "Min": "Min",
     "Minimise": "Réduire",
     "Missed": "Manquée",
     "Mode": "Mode",
@@ -1133,7 +1133,6 @@ const LEMBAS_I18N = {
     "Models": "Modèles",
     "Models for scheduled tasks": "Modèles pour les tâches planifiées",
     "More": "Plus",
-    "Most": "Au plus",
     "Never": "Jamais",
     "New agent chat": "Nouvelle discussion d'agent",
     "New chat": "Nouvelle conversation",
@@ -1256,6 +1255,7 @@ const LEMBAS_I18N = {
     "The model's own": "Celle du modèle",
     "The most an agent chat may do here": "Le maximum qu'une conversation d'agent peut faire ici",
     "The next runs": "Les prochaines exécutions",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Les nombres comptent l’agent de la conversation : 1, c’est l’agent seul, sans sous-agents ; Max 4, c’est l’agent et jusqu’à 3 sous-agents. Un sous-agent utilise le modèle de la conversation, ou un autre seulement si vous le nommez dans la conversation.",
     "The reply is ready.": "La réponse est prête.",
     "The result is no longer there.": "Le résultat n'est plus là.",
     "The server refused this computer's link: %(why)s": "Le serveur a refusé la liaison de cet ordinateur : %(why)s",
@@ -1399,7 +1399,6 @@ const LEMBAS_I18N = {
     "Copied": "Másolva",
     "Copy": "Másolás",
     "Counted by the model at the last reply; the categories are estimated.": "A modell számolta az utolsó válasznál; a kategóriák becslések.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "A beszélgetés saját ügynöke is beleszámít. Az alügynök a beszélgetés modelljén fut, másikon csak akkor, ha megnevezi a beszélgetésben.",
     "Cut short here; the whole of it is in the web interface.": "Itt rövidítve; a teljes a webes felületen van.",
     "Days": "Napok",
     "Delete": "Törlés",
@@ -1425,7 +1424,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Minden eredmény, akkor is, ha nincs mit jelenteni.",
     "Every weekday at 8:00, summarise the news on my reading list": "Minden hétköznap 8:00-kor foglald össze az olvasólistám híreit",
     "Failed": "Sikertelen",
-    "Fewest": "Legalább",
     "Files": "Fájlok",
     "Finished %(n)s": "Kész: %(n)s",
     "Folder": "Mappa",
@@ -1450,11 +1448,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "LLeMbas-kiszolgáló keresése ott…",
     "MCP tools": "MCP-eszközök",
     "Manual — asks first": "Manual — előbb kérdez",
+    "Max": "Max.",
     "Maximise": "Teljes méret",
     "Memory and personal": "Memória és személyes adatok",
     "Menu": "Menü",
     "Message": "Üzenet",
     "Messages": "Üzenetek",
+    "Min": "Min.",
     "Minimise": "Kis méret",
     "Missed": "Elmaradt",
     "Mode": "Mód",
@@ -1462,7 +1462,6 @@ const LEMBAS_I18N = {
     "Models": "Modellek",
     "Models for scheduled tasks": "Modellek az ütemezett feladatokhoz",
     "More": "Több",
-    "Most": "Legfeljebb",
     "Never": "Soha",
     "New agent chat": "Új ügynökcsevegés",
     "New chat": "Új csevegés",
@@ -1585,6 +1584,7 @@ const LEMBAS_I18N = {
     "The model's own": "A modell sajátja",
     "The most an agent chat may do here": "A legtöbb, amit egy ügynök-csevegés itt tehet",
     "The next runs": "A következő futások",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "A számokba a beszélgetés saját ügynöke is beleszámít: az 1 maga az ügynök, alügynökök nélkül; a Max. 4 az ügynök és legfeljebb 3 alügynök. Az alügynök a beszélgetés modelljén fut, másikon csak akkor, ha megnevezi a beszélgetésben.",
     "The reply is ready.": "A válasz kész.",
     "The result is no longer there.": "Az eredmény már nincs meg.",
     "The server refused this computer's link: %(why)s": "A szerver elutasította ennek a számítógépnek a kapcsolatát: %(why)s",
@@ -1728,7 +1728,6 @@ const LEMBAS_I18N = {
     "Copied": "Copiato",
     "Copy": "Copia",
     "Counted by the model at the last reply; the categories are estimated.": "Contato dal modello all’ultima risposta; le categorie sono stime.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Conta anche l’agente della chat. Un sottoagente usa il modello della chat, o un altro solo se lo nomini nella chat.",
     "Cut short here; the whole of it is in the web interface.": "Troncato qui; il resto è nell'interfaccia web.",
     "Days": "Giorni",
     "Delete": "Elimina",
@@ -1754,7 +1753,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Ogni risultato, anche senza nulla da segnalare.",
     "Every weekday at 8:00, summarise the news on my reading list": "Ogni giorno feriale alle 8:00, riassumi le notizie della mia lista di lettura",
     "Failed": "Non riusciti",
-    "Fewest": "Minimo",
     "Files": "File",
     "Finished %(n)s": "Terminate: %(n)s",
     "Folder": "Cartella",
@@ -1779,11 +1777,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Cerco un server LLeMbas lì…",
     "MCP tools": "Strumenti MCP",
     "Manual — asks first": "Manual — chiede prima",
+    "Max": "Max",
     "Maximise": "Ingrandisci",
     "Memory and personal": "Memoria e dati personali",
     "Menu": "Menu",
     "Message": "Messaggio",
     "Messages": "Messaggi",
+    "Min": "Min",
     "Minimise": "Riduci a icona",
     "Missed": "Mancata",
     "Mode": "Modalità",
@@ -1791,7 +1791,6 @@ const LEMBAS_I18N = {
     "Models": "Modelli",
     "Models for scheduled tasks": "Modelli per le attività pianificate",
     "More": "Altro",
-    "Most": "Massimo",
     "Never": "Mai",
     "New agent chat": "Nuova chat con agente",
     "New chat": "Nuova chat",
@@ -1914,6 +1913,7 @@ const LEMBAS_I18N = {
     "The model's own": "Quello del modello",
     "The most an agent chat may do here": "Il massimo che una chat agente può fare qui",
     "The next runs": "Le prossime esecuzioni",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "I numeri contano anche l’agente della chat: 1 è l’agente da solo, senza sottoagenti; Max 4 è l’agente più fino a 3 sottoagenti. Un sottoagente usa il modello della chat, o un altro solo se lo nomini nella chat.",
     "The reply is ready.": "La risposta è pronta.",
     "The result is no longer there.": "Il risultato non c’è più.",
     "The server refused this computer's link: %(why)s": "Il server ha rifiutato il collegamento di questo computer: %(why)s",
@@ -2057,7 +2057,6 @@ const LEMBAS_I18N = {
     "Copied": "Skopiowano",
     "Copy": "Kopiuj",
     "Counted by the model at the last reply; the categories are estimated.": "Policzone przez model przy ostatniej odpowiedzi; kategorie są szacunkowe.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Liczony razem z własnym agentem czatu. Podagent działa na modelu czatu, na innym tylko wtedy, gdy wskażesz go w czacie.",
     "Cut short here; the whole of it is in the web interface.": "Tu skrócone; całość jest w interfejsie WWW.",
     "Days": "Dni",
     "Delete": "Usuń",
@@ -2083,7 +2082,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Każdy wynik, nawet bez niczego do zgłoszenia.",
     "Every weekday at 8:00, summarise the news on my reading list": "W każdy dzień roboczy o 8:00 streść wiadomości z mojej listy do czytania",
     "Failed": "Nieudane",
-    "Fewest": "Najmniej",
     "Files": "Pliki",
     "Finished %(n)s": "Zakończone: %(n)s",
     "Folder": "Folder",
@@ -2108,11 +2106,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Szukam tam serwera LLeMbas…",
     "MCP tools": "Narzędzia MCP",
     "Manual — asks first": "Manual — najpierw pyta",
+    "Max": "Maks",
     "Maximise": "Maksymalizuj",
     "Memory and personal": "Pamięć i dane osobiste",
     "Menu": "Menu",
     "Message": "Wiadomość",
     "Messages": "Wiadomości",
+    "Min": "Min",
     "Minimise": "Minimalizuj",
     "Missed": "Przegapiony",
     "Mode": "Tryb",
@@ -2120,7 +2120,6 @@ const LEMBAS_I18N = {
     "Models": "Modele",
     "Models for scheduled tasks": "Modele dla zaplanowanych zadań",
     "More": "Więcej",
-    "Most": "Najwięcej",
     "Never": "Nigdy",
     "New agent chat": "Nowy czat z agentem",
     "New chat": "Nowy czat",
@@ -2243,6 +2242,7 @@ const LEMBAS_I18N = {
     "The model's own": "Domyślne modelu",
     "The most an agent chat may do here": "Najwięcej, co czat agenta może tu zrobić",
     "The next runs": "Następne uruchomienia",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Liczby obejmują własnego agenta czatu: 1 to sam agent, bez podagentów; Maks 4 to agent i do 3 podagentów. Podagent działa na modelu czatu, na innym tylko wtedy, gdy wskażesz go w czacie.",
     "The reply is ready.": "Odpowiedź jest gotowa.",
     "The result is no longer there.": "Wyniku już nie ma.",
     "The server refused this computer's link: %(why)s": "Serwer odrzucił połączenie tego komputera: %(why)s",
@@ -2386,7 +2386,6 @@ const LEMBAS_I18N = {
     "Copied": "Skopírované",
     "Copy": "Kopírovať",
     "Counted by the model at the last reply; the categories are estimated.": "Spočítal model pri poslednej odpovedi; kategórie sú odhadnuté.",
-    "Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Počíta sa vrátane vlastného agenta konverzácie. Podagent beží na modeli konverzácie, na inom len vtedy, keď ho v konverzácii pomenujete.",
     "Cut short here; the whole of it is in the web interface.": "Tu skrátené; celé je to vo webovom rozhraní.",
     "Days": "Dni",
     "Delete": "Zmazať",
@@ -2412,7 +2411,6 @@ const LEMBAS_I18N = {
     "Every result, even with nothing to report.": "Každý výsledok, aj keď nie je čo hlásiť.",
     "Every weekday at 8:00, summarise the news on my reading list": "Každý pracovný deň o 8:00 zhrň správy z môjho zoznamu na čítanie",
     "Failed": "Zlyhalo",
-    "Fewest": "Najmenej",
     "Files": "Súbory",
     "Finished %(n)s": "Dokončené: %(n)s",
     "Folder": "Priečinok",
@@ -2437,11 +2435,13 @@ const LEMBAS_I18N = {
     "Looking for a LLeMbas server there…": "Hľadám tam server LLeMbas…",
     "MCP tools": "Nástroje MCP",
     "Manual — asks first": "Manual — najprv sa opýta",
+    "Max": "Max",
     "Maximise": "Maximalizovať",
     "Memory and personal": "Pamäť a osobné",
     "Menu": "Ponuka",
     "Message": "Správa",
     "Messages": "Správy",
+    "Min": "Min",
     "Minimise": "Minimalizovať",
     "Missed": "Zmeškané",
     "Mode": "Režim",
@@ -2449,7 +2449,6 @@ const LEMBAS_I18N = {
     "Models": "Modely",
     "Models for scheduled tasks": "Modely pre naplánované úlohy",
     "More": "Viac",
-    "Most": "Najviac",
     "Never": "Nikdy",
     "New agent chat": "Nová agentská konverzácia",
     "New chat": "Nová konverzácia",
@@ -2572,6 +2571,7 @@ const LEMBAS_I18N = {
     "The model's own": "Podľa modelu",
     "The most an agent chat may do here": "Najviac, čo tu agentská konverzácia smie",
     "The next runs": "Ďalšie spustenia",
+    "The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.": "Čísla počítajú aj vlastného agenta konverzácie: 1 je agent sám, bez podagentov; Max 4 je agent a až 3 podagenti. Podagent beží na modeli konverzácie, na inom len vtedy, keď ho v konverzácii pomenujete.",
     "The reply is ready.": "Odpoveď je hotová.",
     "The result is no longer there.": "Výsledok už tu nie je.",
     "The server refused this computer's link: %(why)s": "Server odmietol prepojenie tohto počítača: %(why)s",

ui/js/chat.js

@@ -652,7 +652,7 @@ function effortChip(model, value, onChange) {
 }
 
 /**
- * The subagents a chat may run: a chip saying the fewest and the most, opening the two fields.
+ * The subagents a chat may run: a chip saying Min–Max, opening one range with two handles.
  * `value` { min, max } or empty; `cap` the instance's most, where the server says it.
  */
 function agentsChip(value, onChange, { cap, own } = {}) {
@@ -669,8 +669,7 @@ function agentsChip(value, onChange, { cap, own } = {}) {
         h(
           "div",
           { class: "pop__body" },
-          h("label", { class: "pop__title", for: "agents-min", text: t("Subagents") }),
-          h("div", { class: "agents__heads", "aria-hidden": "true" }, h("span", { text: t("Fewest") }), h("span"), h("span", { text: t("Most") })),
+          h("p", { class: "pop__title", text: t("Subagents") }),
           fields,
           own ? h("p", { class: "hint is-warn", text: SUBAGENTS_OWN() }) : null,
           h("p", { class: "hint", text: SUBAGENTS_HINT() }),

ui/js/models.js

@@ -226,64 +226,87 @@ export function subagentsCap(...sources) {
   }
   return knownCap || SUBAGENTS_CAP
 }
-export const SUBAGENTS_HINT = () => t("Counted with the chat's own agent. A subagent runs on the chat's model, or on another only when you name it in the chat.")
+// The numbers count the chat's own agent: 1 is the agent alone.
+export const SUBAGENTS_HINT = () => t("The numbers count this chat's own agent: 1 is the agent alone, with no subagents; Max 4 is the agent and up to 3 subagents. A subagent runs on the chat's model, or on another only when you name it in the chat.")
 // The spec's `subagents.setting.own`: beside the setting where the chat's model cannot serve them.
 export const SUBAGENTS_OWN = () => t("This model serves one request at a time: subagents need another model, on another server, named in the chat.")
 
+// A new chat's setting as the harness spec seeds it (`loop.json` agents): 1 to 4.
+export const AGENTS_DEFAULT = { min: 1, max: 4 }
+
 /**
- * The subagents a chat may run: the fewest and the most, as two number fields on one shared
- * track. `value` is { min, max } (empty fields for a new chat: the server's defaults then stand);
- * `cap` the instance's most; `onChange({ min, max })` once a field is changed, with the empty one
- * left out. The server checks the pair and says the bounds when it refuses one.
+ * The agents a chat may run, its own agent counted: one range from 1 to `cap` with two handles,
+ * Min and Max -- two range inputs overlaid on one track, the span between them filled. Each is a
+ * real slider (arrow keys, Home/End, Page Up/Down), and Min never passes Max. `value` is
+ * { min, max } (empty for a new chat: the spec's defaults are shown, and left to the server
+ * unless a handle moves); `el.value` the pair. The server checks it again.
  */
-export function agentsControl({ value, onChange, id, compact, cap = SUBAGENTS_CAP }) {
-  const field = (which, v, label) =>
+export function agentsControl({ value, onChange, id, cap = SUBAGENTS_CAP }) {
+  const top = Math.max(1, Number(cap) || SUBAGENTS_CAP)
+  const clamp = (n, d) => Math.min(top, Math.max(1, Number.isFinite(Number(n)) && n !== null && n !== "" ? Number(n) : d))
+  let given = value || {}
+  let touched = false
+  const handle = (which, label) =>
     h("input", {
-      type: "number",
-      class: "agents__input",
+      type: "range",
+      class: `range2__input range2__input--${which}`,
       id: which === "min" ? id : null,
       min: "1",
-      max: String(cap),
+      max: String(top),
       step: "1",
-      inputmode: "numeric",
-      value: v === undefined || v === null ? "" : String(v),
-      placeholder: compact ? null : t("default"),
       "aria-label": label,
       title: label,
     })
-  const low = field("min", value?.min, t("Fewest"))
-  const high = field("max", value?.max, t("Most"))
-  const bounds = () => {
-    high.min = low.value || "1"
-    low.max = high.value || String(cap)
+  const low = handle("min", t("Min"))
+  const high = handle("max", t("Max"))
+  const said = h("output", { class: "range2__said", "aria-live": "polite", for: id || null })
+  const fill = h("span", { class: "range2__fill", "aria-hidden": "true" })
+  const box = h("span", { class: "range2__box" }, h("span", { class: "range2__track", "aria-hidden": "true" }, fill), low, high)
+  const draw = () => {
+    const lo = Number(low.value)
+    const hi = Number(high.value)
+    const at = (n) => (top > 1 ? ((n - 1) / (top - 1)) * 100 : 0)
+    box.style.setProperty("--lo", `${at(lo)}%`)
+    box.style.setProperty("--hi", `${at(hi)}%`)
+    // Where the two meet, the one that can still move is on top: Min at the far end, else Max.
+    low.classList.toggle("is-top", lo === hi && hi === top)
+    low.setAttribute("aria-valuetext", `${t("Min")} ${lo}`)
+    high.setAttribute("aria-valuetext", `${t("Max")} ${hi}`)
+    said.textContent = `${t("Min")} ${lo} · ${t("Max")} ${hi}`
   }
-  bounds()
-  const read = () => {
-    const out = {}
-    if (low.value) out.min = Number(low.value)
-    if (high.value) out.max = Number(high.value)
-    return out
+  const set = (v) => {
+    const hi = clamp(v?.max, Math.min(AGENTS_DEFAULT.max, top))
+    const lo = Math.min(hi, clamp(v?.min, AGENTS_DEFAULT.min))
+    high.value = String(hi)
+    low.value = String(lo)
+    draw()
   }
-  for (const input of [low, high]) {
-    input.addEventListener("change", () => {
-      bounds()
-      onChange?.(read())
+  low.addEventListener("input", () => {
+    if (Number(low.value) > Number(high.value)) low.value = high.value
+    touched = true
+    draw()
   })
-  }
+  high.addEventListener("input", () => {
+    if (Number(high.value) < Number(low.value)) high.value = low.value
+    touched = true
+    draw()
+  })
+  for (const input of [low, high]) input.addEventListener("change", () => onChange?.(read()))
+  const read = () => (touched || given.min || given.max ? { min: Number(low.value), max: Number(high.value) } : {})
+  set(given)
   const el = h(
     "span",
-    { class: `agents${compact ? " agents--compact" : ""}`, role: "group", "aria-label": t("Subagents"), title: SUBAGENTS_HINT() },
-    compact ? h("span", { class: "agents__label", text: t("Subagents") }) : null,
-    low,
-    h("span", { class: "agents__dash", "aria-hidden": "true", text: "–" }),
-    high,
+    { class: "range2", role: "group", "aria-label": t("Subagents") },
+    said,
+    box,
+    h("span", { class: "range2__ends", "aria-hidden": "true" }, h("span", { text: "1" }), h("span", { text: String(top) })),
   )
   Object.defineProperty(el, "value", {
     get: read,
     set: (v) => {
-      low.value = v?.min ?? ""
-      high.value = v?.max ?? ""
-      bounds()
+      given = v || {}
+      touched = false
+      set(given)
     },
   })
   return el