A crowd that does not agree with whoever spoke last

Three fixes to how a round behaves, found by reading one real round on the live
instance rather than by testing it.

A member asked "what would you have done differently" answered the person's
original question again instead of critiquing what was already there. Fine on a
question with one answer; on a request to *make* something it is an invitation.
`crowd.turn` now says to respond to what is above and not to re-answer.

The model that opened the round, told to write the final answer and take what
the others got right, abandoned its own good answer and adopted the newcomer's
position with no argument anywhere for why. Both closing fragments now say that
an answer is not the worse one for having been written first, and that agreement
with no argument behind it is not a reason to change.

That second one is not cosmetic: all three answers from the observed round were
compiled. The original and the critic's alternative both build; the merged
answer that was actually delivered does not. A crowd's failure mode is not
looping -- the caps handle that -- it is converging on the last thing said.

Third, the reply that opens a round now carries a chip like every other one. It
is the single contribution the crowd does not start, so there was nothing to
stamp it with until the round began, and a two-model round rendered as an
unmarked reply followed by one saying "2 of 2". The stamp is display state and
never scheduling state: `crowd.scheduling_state` hides it from everything that
decides what happens next, because fed to the scheduler it would inherit the
round's clock -- regenerating the opening an hour later would end the round with
"out of time" before anybody spoke -- and would hand that reply a member's tools
and a member's instruction.

And the chip was never translated. It is now, with the count as placeholders
rather than three t() calls around one sentence.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

CHANGELOG.md

@@ -16,6 +16,88 @@ for 1.0.0 have something to be assembled from.
 
 ## Unreleased
 
+## 1.8.1
+
+Three fixes to how a crowd behaves, found by reading one real round on the live
+instance rather than by testing: two models, one round, a question that asked for
+something to be *made*.
+
+- **A member no longer answers the question again.** Asked to pick a language and
+  write an example, the main model wrote Python; the second model gave a genuinely
+  useful critique of it — and then answered the original question itself, in a
+  different language. Nothing in its instruction said not to. It now says so:
+  *respond to what is above you; do not answer the person's original request again
+  yourself.* A member that produces a rival answer is not a second opinion, it is
+  a second first opinion, and it is what takes a round off the question.
+
+- **The model that opened the round no longer capitulates.** Told to write the
+  final answer and take what the others got right, it abandoned its own perfectly
+  good answer, wrote *"I agree that Rust is the superior choice"* with no argument
+  anywhere for why, and rewrote everything in the newcomer's language. Both
+  closing instructions now carry: *your own answer is not automatically the worse
+  one for having been written first; change your position where somebody gave you
+  a reason, and say what the reason was.*
+
+  This mattered more than it reads. All three answers were compiled: the original
+  Python was fine, the critic's Rust compiled and ran — and **the merged answer
+  that was actually delivered did not compile at all**. A crowd that ends by
+  agreeing with whoever spoke last can be worse than the model that started it.
+
+- **The bubble that opens a round now says `1 of 3` like every other one.** It was
+  the single contribution with no chip, because the crowd does not start it — the
+  composer does, and a round only begins when it finishes. So a two-model round
+  read as an ordinary reply followed by one labelled `2 of 2`, with no 1 anywhere.
+  It is stamped when the round begins, and that stamp is deliberately invisible to
+  everything that decides what happens next: fed to the scheduler it would inherit
+  the round's clock, so regenerating the opening an hour later would end the round
+  with "out of time" before anybody spoke.
+
+- Fixed: **the crowd chip was never translated.** `1 of 3`, `on the way back`,
+  `closing`, `no rounds left` and the rest were English on a Slovak instance.
+
+**Worth knowing, and not a bug:** with **two** models there is no backward pass at
+all. The way back would contain only the model that opened the round, whose turn
+*is* the close — so `crowd.disagree` never fires. You need at least three models
+before a single "do you disagree" bubble can exist.
+
+## 1.8.0
+
+- **The crowd is where you would look for it.** In 1.6.0 the only way to add a
+  model to a chat was the Chat settings panel — behind the ⋯ menu, inside a chat
+  that already existed — and the switch that turns the feature on was a card on
+  the Agents page. Somebody who enabled it went looking and found nothing, which
+  is the correct outcome of that arrangement.
+
+  Now there is a **crowd button in the composer**, beside the attachment and
+  scope buttons, on both the chat screen and Messages. It carries a count when
+  the chat has a crowd, it lists the models you can reach, and it says what the
+  turn will cost before you tick anything. On the new-chat screen the choice
+  **rides along with the first message**, so a chat can start as a crowd rather
+  than having to be converted into one.
+
+  The instance switch and its bounds have moved to their own page, **Admin →
+  Crowd**.
+
+- Fixed: **the new-chat screen was wider than a phone.** Before the first
+  message, the suggestion cards pushed the conversation 65px past the edge of a
+  390px screen and it could be dragged sideways; after the first message it
+  looked right, because the cards were gone. Reported from a phone.
+
+  Two things were true at once. The cards' grid asked for a minimum column width
+  it could not give up — the ordinary version of this bug — and it was *also* a
+  grid item, which means it carried a min-content floor that beats `width: 100%`
+  outright. Fixing only the first made it 27px worse. Both are fixed, on all four
+  grids in the stylesheets that could have it, and a test now refuses either half
+  of the pair on its own.
+
+  The reason this survived four releases of narrow-width checking is worth
+  recording: the screenshot harness built its client without running the
+  application's startup, so the suggestion cards were **absent from every shot
+  ever taken of that screen**, and its overflow check deliberately ignored
+  anything inside a scrolling box — correct for a wide table in its own scroller,
+  blind to a box that scrolls sideways when nobody asked it to. Both are fixed,
+  and the harness now names the offending element and the child responsible.
+
 ## 1.7.0
 
 - **The interface speaks Slovak.** Pick a language under **Appearance** in your

scripts/shoot.py

@@ -26,6 +26,7 @@ import shutil
 import subprocess
 import sys
 import tempfile
+from functools import cache
 from pathlib import Path
 
 REPO = Path(__file__).resolve().parent.parent
@@ -128,8 +129,65 @@ window.__measure = function () {
     return found.sort(function (a, b) { return b.over - a.over; }).slice(0, 8);
   }
 
+  /* --- A box that scrolls sideways when nobody asked it to -----------------
+
+     The blind spot that hid the suggestions bug through forty measurements.
+     `.suggestions` rendered 455px wide inside a 390px `.thread-scroll`, and
+     every check above looked straight past it: `culprits('x')` skips anything
+     with a scrollable ancestor -- correct for a table inside its own scroller,
+     wrong for the scroller itself -- and `scrollsSideways` stayed false because
+     `.thread-scroll` absorbed the overflow instead of the document.
+
+     "Authored" is the distinction that makes this reportable rather than noise.
+     The tree's rule is that anything wide gets its OWN scroller, so a wrapper
+     carrying `overflow-x: auto` in a stylesheet is right. A box given only
+     `overflow-y: auto` scrolls sideways as well, because the other axis then
+     computes to `auto` -- and that is always a bug. Computed style cannot tell
+     those apart, both being `auto`, so the rules that say it are read off the
+     stylesheets -- in Python, by `authored_sideways()` below, and not from the
+     CSSOM here: a stylesheet loaded over `file://` is a foreign origin for
+     `cssRules` even with `--allow-file-access-from-files`, and every sheet
+     throws. That silently found *nothing authored*, which turns this check into
+     "every vertical scroller is a bug" -- so the list arriving empty is a hard
+     error rather than a clean run. */
+  var sidewaysAuthors = __SIDEWAYS_AUTHORS__;
+
+  function authoredSideways(el) {
+    if (el.style.overflowX || el.style.overflow) return true;
+    for (var i = 0; i < sidewaysAuthors.length; i++) {
+      try { if (el.matches(sidewaysAuthors[i])) return true; } catch (e) { /* :has() etc */ }
+    }
+    return false;
+  }
+
+  var sideways = [];
+  document.querySelectorAll('body, body *').forEach(function (el) {
+    var ox = getComputedStyle(el).overflowX;
+    if (ox !== 'auto' && ox !== 'scroll') return;
+    if (el.scrollWidth <= el.clientWidth + 1) return;
+    if (authoredSideways(el)) return;
+    /* Which child is doing it. "`.thread-scroll` scrolls sideways" is not
+       actionable; "`.suggestions` is 455px inside its 390px" is. */
+    var worst = null;
+    el.querySelectorAll('*').forEach(function (kid) {
+      var over = kid.getBoundingClientRect().width - el.clientWidth;
+      if (over > 1 && (!worst || over > worst.over)) {
+        worst = {tag: kid.tagName.toLowerCase(),
+                 cls: (kid.className && kid.className.toString().slice(0, 50)) || '',
+                 w: Math.round(kid.getBoundingClientRect().width),
+                 over: Math.round(over)};
+      }
+    });
+    sideways.push({tag: el.tagName.toLowerCase(),
+                   cls: (el.className && el.className.toString().slice(0, 50)) || '',
+                   scrollW: el.scrollWidth, clientW: el.clientWidth,
+                   widest: worst});
+  });
+
   var shell = document.querySelector('.shell');
   return {
+    sidewaysScrollers: sideways.slice(0, 8),
+    sidewaysCount: sideways.length,
     docScrollH: de.scrollHeight,
     innerH: window.innerHeight,
     docScrollW: de.scrollWidth,
@@ -167,6 +225,42 @@ window.__measure = function () {
 """
 
 
+@cache
+def authored_sideways() -> tuple[str, ...]:
+    """Selectors whose rules really do ask for horizontal scrolling.
+
+    The tree's rule is that anything wide gets its own scroller, so these are
+    the correct ones: a table wrapper, a code block, the tab bar. Everything
+    else that scrolls sideways is `overflow-y: auto` dragging the other axis
+    along with it, which is always a bug and is what `.suggestions` did.
+    """
+    selectors: list[str] = []
+    for path in sorted((STATIC / "css").glob("*.css")):
+        text = re.sub(r"/\*.*?\*/", "", path.read_text(), flags=re.S)
+        # Innermost blocks only: `[^{}]*` cannot cross a brace, so an `@media`
+        # prelude never matches and the rules inside it do.
+        for prelude, body in re.findall(r"([^{}]*)\{([^{}]*)\}", text):
+            wants = False
+            for declaration in body.split(";"):
+                name, _, value = declaration.partition(":")
+                name, value = name.strip().lower(), value.strip().lower()
+                if name not in ("overflow", "overflow-x") or not value:
+                    continue
+                # `overflow: hidden auto` is x then y, so the first word is ours;
+                # `overflow: auto` is both.
+                wants = wants or value.split()[0] in ("auto", "scroll")
+            if not wants:
+                continue
+            selectors += [
+                part.strip()
+                for part in prelude.split(",")
+                if part.strip() and not part.strip().startswith("@")
+            ]
+    if not selectors:
+        raise SystemExit("read no horizontal-overflow rules -- the sideways check would cry wolf")
+    return tuple(selectors)
+
+
 def build_client():
     import lembas.config as config_mod
 
@@ -198,6 +292,17 @@ def build_client():
         db.flush()
         for name in ("gemma4-moe", "qwen3-coder"):
             db.add(Model(connection_id=connection.id, model_id=name, display_name=name))
+
+    # 🚨 The suggestion cards are seeded by the startup hook, and `TestClient(app)`
+    # runs a lifespan only inside a `with` block -- so every shot of the new-chat
+    # screen ever taken by this script was of a page with its cards missing. That
+    # is how a grid 65px wider than a phone survived forty measurements. Seeded
+    # here rather than by entering the lifespan, which would also start the
+    # schedule ticker and rehydrate background jobs inside a screenshot run.
+    from lembas.services.suggestions import seed_defaults as seed_suggestions
+
+    with session_scope() as db:
+        seed_suggestions(db)
     return client
 
 
@@ -219,6 +324,20 @@ def rewrite(html: str, client, assets: Path) -> str:
         html,
     )
 
+    # Anything else the *application* serves rather than mounts. Model avatars live
+    # under `/uploads/models/…`, which is a route behind auth -- so they cannot be
+    # pointed at a file on disk and have to be fetched through the client like
+    # `/branding.css` above. A real instance has them and a fixture does not, which
+    # is exactly the difference that makes a page measured here unlike the page
+    # somebody is looking at.
+    for url in sorted({*re.findall(r'\bsrc="(/(?:uploads|branding)/[^"?]+)"', html)}):
+        response = client.get(url)
+        if response.status_code != 200:
+            continue
+        name = "fetched-" + url.strip("/").replace("/", "-")
+        (assets / name).write_bytes(response.content)
+        html = html.replace(f'src="{url}"', f'src="file://{assets}/{name}"')
+
     # Fail loudly, and only about things that decide how the page LOOKS: every
     # `src`, and `href` on a <link>. An `href` on an anchor is a destination,
     # not an asset -- flagging those makes the guard cry wolf on every page and
@@ -249,14 +368,15 @@ def rewrite(html: str, client, assets: Path) -> str:
     if missing:
         raise SystemExit(f"REWRITTEN TO NOTHING -- still an unstyled document: {missing[:5]}")
 
-        # The one-time notifications offer is a modal over the very page we came
+    # The one-time notifications offer is a modal over the very page we came
     # to measure, and it is gated on a localStorage key. Set it in the head, so
     # it runs before the deferred script that reads it.
     quiet = (
         "<script>try{localStorage.setItem('lembas-notifications-asked','1');}"
         "catch(e){}</script>"
     )
-    return html.replace("</head>", quiet + MEASURE + "</head>", 1)
+    measure = MEASURE.replace("__SIDEWAYS_AUTHORS__", json.dumps(list(authored_sideways())))
+    return html.replace("</head>", quiet + measure + "</head>", 1)
 
 
 def shoot(client, path: str, width: int, height: int, theme: str, outdir: Path) -> dict:
@@ -391,6 +511,13 @@ def main() -> None:
             flags.append(f"DOC-SCROLLS({r['docScrollH']}>{r['innerH']})")
         if r["scrollsSideways"]:
             flags.append(f"SIDEWAYS({r['docScrollW']}>{r['innerW']})")
+        for s in r.get("sidewaysScrollers", []):
+            widest = s["widest"]
+            blame = f"<{widest['tag']}.{widest['cls']} {widest['w']}px" if widest else ""
+            flags.append(
+                f"SCROLLER-SIDEWAYS({s['tag']}.{s['cls']} "
+                f"{s['scrollW']}>{s['clientW']}{blame})"
+            )
         if r["overflowCount"]:
             flags.append(f"overflow:{r['overflowCount']}")
         if r["smallCount"]:

src/lembas/__init__.py

@@ -1,3 +1,3 @@
 """LLeMbas - a Middle-earth themed web UI for OpenAI-compatible LLM endpoints."""
 
-__version__ = "1.7.0"
+__version__ = "1.8.1"

src/lembas/api/admin_agents.py

@@ -66,10 +66,6 @@ async def agents_page(request: Request, db: Db, user: AdminUser, saved: bool = F
             # reply is allowed to set going on its own, and a nav entry for one
             # card would be worse than the near-miss.
             "subagents": settings_store.subagents(db),
-            # And a third group on the same page, for the same reason: a crowd is
-            # not an agent-chat feature either, but this is where somebody comes to
-            # find out what one turn is allowed to set going.
-            "crowd": settings_store.crowd(db),
             "saved": saved,
         },
     )
@@ -115,35 +111,6 @@ async def save_subagents(
     return RedirectResponse("/admin/agents?saved=1", status_code=status.HTTP_303_SEE_OTHER)
 
 
-@router.post("/crowd")
-async def save_crowd(
-    db: Db,
-    user: AdminUser,
-    enabled: bool = Form(False),
-    max_models: int = Form(4),
-    max_rounds: int = Form(2),
-    wall_seconds: int = Form(900),
-    collapse_agreement: bool = Form(False),
-) -> Response:
-    """Its own route, for the reason `save_subagents` gives above."""
-    settings_store.update(
-        db,
-        {
-            "enabled": enabled,
-            # Clamped here as well as on read. Every floor is one: a zero would be
-            # the feature switched off wearing the switch's clothes, and that is a
-            # thing to answer in one place.
-            "max_models": min(max(max_models, 1), 8),
-            "max_rounds": min(max(max_rounds, 1), 5),
-            "wall_seconds": min(max(wall_seconds, 60), 7200),
-            "collapse_agreement": collapse_agreement,
-        },
-        key=settings_store.CROWD,
-    )
-    log.info("crowd %s by %s", "enabled" if enabled else "disabled", user.email)
-    return RedirectResponse("/admin/agents?saved=1", status_code=status.HTTP_303_SEE_OTHER)
-
-
 @router.post("")
 async def save_agents(
     db: Db,

src/lembas/api/admin_crowd.py

@@ -0,0 +1,68 @@
+"""The crowd: several models answering one turn, in any chat.
+
+Its own module because it is its own page, and it is its own page because as a card
+on `/admin/agents` it read as an agent-chat feature. It is not one: a crowd works in
+an ordinary conversation, and the owner reasonably concluded otherwise from where
+the switch was sitting.
+"""
+
+from __future__ import annotations
+
+import logging
+
+from fastapi import APIRouter, Form, Request, Response, status
+from fastapi.responses import RedirectResponse
+
+from lembas.api.deps import AdminUser, Db
+from lembas.services import settings_store
+from lembas.web.templating import render
+
+log = logging.getLogger(__name__)
+
+router = APIRouter(prefix="/admin/crowd", tags=["admin-crowd"])
+
+
+@router.get("")
+async def crowd_page(request: Request, db: Db, user: AdminUser, saved: str = ""):
+    """Its own page, for the reason its template records: as a card on the Agents
+    screen it read as an agent-chat feature, which it is not."""
+    return render(
+        request,
+        "admin/crowd.html",
+        {"crowd": settings_store.crowd(db), "saved": saved},
+    )
+
+
+@router.post("")
+async def save_crowd(
+    db: Db,
+    user: AdminUser,
+    enabled: bool = Form(False),
+    max_models: int = Form(4),
+    max_rounds: int = Form(2),
+    wall_seconds: int = Form(900),
+    collapse_agreement: bool = Form(False),
+) -> Response:
+    """One group, one form, one route.
+
+    The bounds are clamped here as well as in `settings_store.crowd`, which is the
+    same belt-and-braces `save_subagents` in `admin_agents.py` uses: a value posted
+    past this route -- by an older page, or by hand -- still reads back sane.
+    """
+    settings_store.update(
+        db,
+        {
+            "enabled": enabled,
+            # Every floor is one: a zero would be the feature switched off
+            # wearing the switch's clothes.
+            "max_models": min(max(max_models, 1), 8),
+            "max_rounds": min(max(max_rounds, 1), 5),
+            "wall_seconds": min(max(wall_seconds, 60), 7200),
+            "collapse_agreement": collapse_agreement,
+        },
+        key=settings_store.CROWD,
+    )
+    log.info("crowd %s by %s", "enabled" if enabled else "disabled", user.email)
+    return RedirectResponse("/admin/crowd?saved=1", status_code=status.HTTP_303_SEE_OTHER)
+
+

src/lembas/api/chats.py

@@ -296,6 +296,11 @@ async def start_chat(
     scope_on: list[str] = Form(default=[]),
     scope_skill_all: list[str] = Form(default=[]),
     scope_skill_on: list[str] = Form(default=[]),
+    # Who else answers, as the crowd menu stood before the first word. There is no
+    # chat row yet to attach members to, so the choice rides along with the message
+    # -- the same mechanism the scope switches above use, and the reason the control
+    # lives inside the composer's form rather than in the topbar.
+    crowd_model_ids: list[str] = Form(default=[]),
 ) -> Response:
     """Create a chat from its first message.
 
@@ -329,6 +334,8 @@ async def start_chat(
         skills_off=frozenset(scope_skill_all) - frozenset(scope_skill_on),
     )
 
+    _apply_crowd(db, chat, user, crowd_model_ids)
+
     _adopt_draft(db, user, draft_id, chat)
 
     user_message = chat_service.create_message(db, chat, ROLE_USER, content)
@@ -1510,6 +1517,45 @@ def _thread_context(db: DBSession, chat: Chat, user: User) -> dict:
     }
 
 
+def _apply_crowd(db: DBSession, chat: Chat, user: User, values: list[str]) -> None:
+    """Replace a chat's crowd with the models named, in the order named.
+
+    One implementation for both the composer (where the choice rides along with
+    the first message) and the settings panel, because two would be two places to
+    forget a rule -- and there are three:
+
+    * **Checked against what this person can reach**, never against what exists.
+      A control checked only in the template is advisory, and a crafted request
+      walks past it. Same reasoning as the model branch in `update_chat`.
+    * **Never the chat's own model**, which would answer twice in a row.
+    * **Capped by `crowd.max_models`**, on the way in as well as on the way out.
+
+    The connection is stored beside the id because `Model` is unique on the pair,
+    and a model offered by two connections is two rows with different capabilities.
+    """
+    from lembas.db.models import CrowdMember
+
+    settings = settings_store.crowd(db)
+    reachable = {
+        model.model_id: model for model in chat_service.available_models(db, user)
+    }
+    wanted: list[str] = []
+    for value in values:
+        value = str(value).strip()
+        if value and value in reachable and value != chat.model_id and value not in wanted:
+            wanted.append(value)
+    wanted = wanted[: int(settings["max_models"])]
+
+    chat.crowd = [
+        CrowdMember(
+            model_id=model_id,
+            connection_id=reachable[model_id].connection_id,
+            position=index,
+        )
+        for index, model_id in enumerate(wanted)
+    ]
+
+
 def _messages_after(db: DBSession, message: Message) -> list[Message]:
     """Everything later in this chat than one message.
 
@@ -2117,39 +2163,8 @@ async def update_chat(request: Request, db: Db, user: RequiredUser, chat_id: str
 
     if "crowd_model_ids" in form:
         # The same shape as the bases above: one field always sent, so clearing
-        # every box clears the crowd. Checked against what this person can reach
-        # rather than against what exists, or the picker is advisory and a crafted
-        # request walks past it -- the reasoning the model branch carries.
-        from lembas.db.models import CrowdMember
-
-        settings = settings_store.crowd(db)
-        reachable = {
-            model.model_id for model in chat_service.available_models(db, user)
-        }
-        wanted: list[str] = []
-        for value in form.getlist("crowd_model_ids"):
-            value = str(value).strip()
-            # Never the chat's own model: it would answer twice in a row, which is
-            # nobody's idea of a second opinion.
-            if value and value in reachable and value != chat.model_id and value not in wanted:
-                wanted.append(value)
-        wanted = wanted[: int(settings["max_models"])]
-
-        chat.crowd = [
-            CrowdMember(
-                model_id=model_id,
-                connection_id=next(
-                    (
-                        model.connection_id
-                        for model in chat_service.available_models(db, user)
-                        if model.model_id == model_id
-                    ),
-                    None,
-                ),
-                position=index,
-            )
-            for index, model_id in enumerate(wanted)
-        ]
+        # every box clears the crowd.
+        _apply_crowd(db, chat, user, form.getlist("crowd_model_ids"))
 
     submitted_params = {name: form[name] for name in _PARAM_RANGES if name in form}
     if submitted_params:

src/lembas/api/pages.py

@@ -83,7 +83,9 @@ def _chat_context(db: DBSession, user: User, chat: Chat | None) -> dict:
             else []
         ),
         "attached_base_ids": [base.id for base in chat.knowledge_bases] if chat else [],
-        **_crowd_context(db, user, chat, models),
+        **_crowd_context(
+            db, user, chat, models, current.model_id if current is not None else ""
+        ),
         # What *this* model takes, not the three every model used to be assumed
         # to take. The vocabulary is per model -- gpt-oss has no `xhigh` and
         # Bonsai has no `high`, and sending the wrong one does not degrade, it
@@ -197,12 +199,17 @@ def _scope_context(db: DBSession, user: User, chat: Chat | None) -> dict:
     }
 
 
-def _crowd_context(db: DBSession, user: User, chat: Chat | None, models: list) -> dict:
+def _crowd_context(
+    db: DBSession, user: User, chat: Chat | None, models: list, default_model_id: str = ""
+) -> dict:
     """Who else could answer in this chat, and what that would cost.
 
-    Empty — and the panel then shows nothing rather than an empty control — when
-    the feature is off, when there is nobody else to add, or on the new-chat
-    screen, where there is no chat to attach anybody to yet.
+    Offered on the **new-chat screen as well**, where there is no chat row yet: the
+    choice rides along with the first message, the way the scope switches do. The
+    first version of this was per-chat only and therefore invisible to anybody
+    setting a conversation up — which is how the feature shipped switched on and
+    unreachable. Empty only when the feature is off or there is nobody else to add,
+    and then the control is absent rather than being an empty menu.
 
     The cost is spelled out because it is the thing somebody will not have thought
     about: a turn is `speakers x rounds x 2 - 1` replies, and on one local endpoint
@@ -211,21 +218,31 @@ def _crowd_context(db: DBSession, user: User, chat: Chat | None, models: list) -
     from lembas.services import crowd as crowd_service
 
     settings = settings_store.crowd(db)
-    if chat is None or not settings["enabled"]:
+    if not settings["enabled"]:
         return {"crowd_available": [], "crowd_member_ids": [], "crowd_skipped": []}
 
-    others = [model for model in models if model.model_id != chat.model_id]
-    members = [
-        row.model_id
-        for row in sorted(chat.crowd, key=lambda row: (row.position, row.model_id))
-    ]
+    # On the new-chat screen the "own" model is whichever one the picker is
+    # showing, so the list excludes it for the same reason it does in a chat:
+    # adding it would have it answer twice in a row.
+    own = chat.model_id if chat is not None else default_model_id
+    others = [model for model in models if model.model_id != own]
+    members = (
+        [
+            row.model_id
+            for row in sorted(chat.crowd, key=lambda row: (row.position, row.model_id))
+        ]
+        if chat is not None
+        else []
+    )
     reachable = {model.model_id for model in others}
     speakers = 1 + len([model_id for model_id in members if model_id in reachable])
     rounds = int(settings["max_rounds"])
     return {
         "crowd_available": others,
         "crowd_member_ids": [model_id for model_id in members if model_id in reachable],
-        "crowd_skipped": crowd_service.unreachable_members(db, chat, user),
+        "crowd_skipped": (
+            crowd_service.unreachable_members(db, chat, user) if chat is not None else []
+        ),
         # One round is out and back: everybody answers, everybody but the last is
         # asked whether they disagree, and the main model closes.
         "crowd_replies": max(1, speakers * 2 - 1),

src/lembas/main.py

@@ -17,6 +17,7 @@ from lembas.api import (
     admin_agents,
     admin_audio,
     admin_branding,
+    admin_crowd,
     admin_extraction,
     admin_images,
     admin_models,
@@ -221,6 +222,7 @@ def create_app() -> FastAPI:
     app.include_router(admin_suggestions.router)
     app.include_router(admin_tools.router)
     app.include_router(admin_agents.router)
+    app.include_router(admin_crowd.router)
     app.include_router(push.router)
     app.include_router(branding.router)
 

src/lembas/services/chat.py

@@ -588,7 +588,10 @@ def build_request(
     if crowd_turn is None and upto is not None:
         from lembas.services import crowd as crowd_service
 
-        crowd_turn = crowd_service.state_of(upto)
+        # `scheduling_state`: the opening reply carries a stamp for the chip's
+        # sake, and regenerating it must still build an ordinary first answer --
+        # not one told that "the answers above are quoted, yours comes next".
+        crowd_turn = crowd_service.scheduling_state(upto)
     # Images are only sent to a model an administrator has marked as having
     # vision. Sending them to one that has not is not a graceful degradation:
     # most endpoints reject the whole request.

src/lembas/services/crowd.py

@@ -134,6 +134,41 @@ def state_of(message: Message | None) -> Turn | None:
         return None
 
 
+def is_opening(state: Turn | None) -> bool:
+    """Whether this state is the main model's opening reply.
+
+    `phase=out, index=0` is **display state and never scheduling state**. The
+    opening reply is not started by the crowd -- the composer starts it, exactly
+    as it starts every other reply, and a round only begins when it *finishes*.
+    Stamping it afterwards is what lets the transcript say `1 of 3` on the bubble
+    that opened the round; before that it was the one contribution with no chip,
+    so a two-model round read as an ordinary reply followed by a crowd.
+
+    Everything that asks "is a round already in progress?" has to skip it, or the
+    stamp changes behaviour it was never meant to touch -- see `scheduling_state`.
+    """
+    return state is not None and state.phase == PHASE_OUT and state.index == 0
+
+
+def scheduling_state(message: Message | None) -> Turn | None:
+    """The round state the scheduler should act on: `state_of`, minus the opening.
+
+    Two things would break if the opening stamp were fed to `next_turn` as real
+    state, and both are silent:
+
+    * **`started_at` would be inherited on a regenerate.** Regenerating the
+      opening reply an hour later would hand `next_turn` an hour-old clock and the
+      round would stop with "out of time" before anybody spoke.
+    * **The once-per-turn gates key off "no state at all"** -- compaction, the
+      title, the unread push. A stamped opening reads as a later speaker, and each
+      of them would be skipped for the turn that is supposed to have them.
+
+    So the stamp is written where the transcript reads it and nowhere else.
+    """
+    state = state_of(message)
+    return None if is_opening(state) else state
+
+
 def now_stamp() -> str:
     return datetime.now(UTC).isoformat()
 
@@ -374,9 +409,11 @@ __all__ = [
     "Turn",
     "elapsed",
     "is_newest",
+    "is_opening",
     "member_speakers",
     "next_turn",
     "now_stamp",
+    "scheduling_state",
     "state_of",
     "tool_defs",
     "unreachable_members",

src/lembas/services/generation.py

@@ -661,7 +661,10 @@ async def _run(generation: Generation) -> None:
             # once, here, and used for three decisions: which tools it may have,
             # which instruction closes its request, and whether it may ask for
             # another round.
-            crowd_state = crowd_service.state_of(message)
+            # `scheduling_state` for the reason `build_request` gives: the
+            # opening reply's stamp is for the transcript, and regenerating it
+            # must not hand it a member's tools or a member's instruction.
+            crowd_state = crowd_service.scheduling_state(message)
             crowd_settings = settings_store.crowd(db)
             may_ask_again = bool(
                 crowd_state is not None
@@ -2232,7 +2235,11 @@ def _advance_crowd(generation: Generation) -> bool:
             speakers = crowd_service.member_speakers(db, chat, owner_user)
             speakers = speakers[: int(settings["max_models"]) + 1]
 
-            state = crowd_service.state_of(message)
+            # `scheduling_state` and not `state_of`: the opening reply carries a
+            # stamp for the transcript's sake (so it can say `1 of 3`), and that
+            # stamp must not read as "a round is already running" -- it would
+            # inherit the old clock on a regenerate. See `crowd.is_opening`.
+            state = crowd_service.scheduling_state(message)
             # The turn a round belongs to: the user message this all answers.
             turn_id = state.turn if state is not None else _turn_anchor(db, message)
             following = crowd_service.next_turn(
@@ -2254,6 +2261,22 @@ def _advance_crowd(generation: Generation) -> bool:
                 db.commit()
                 return False
 
+            if state is None:
+                # The round begins here, so stamp the reply that opened it. It is
+                # the only contribution that is not started by the crowd, and
+                # before this it was the only one with no chip -- which made a
+                # two-model round read as an ordinary reply followed by a crowd,
+                # and left the reader counting "2 of 2" with no 1 in sight. Same
+                # turn and same `started_at`, so the bubbles group.
+                message.crowd_json = crowd_service.Turn(
+                    turn=following.turn,
+                    round=following.round,
+                    phase=crowd_service.PHASE_OUT,
+                    index=0,
+                    of=following.of,
+                    started_at=following.started_at,
+                ).as_json()
+
             speaker = speakers[following.index]
             placeholder = chat_service.create_message(
                 db,
@@ -2281,10 +2304,14 @@ def _opens_the_turn(message: Message) -> bool:
     """Whether this reply is the first one answering a question.
 
     True for every ordinary reply, and for a crowd only for the main model's
-    opening turn -- which is the one with no crowd state on it at all, because a
-    round begins when that reply *finishes*.
+    opening turn. That reply has no crowd state while it is being written -- a
+    round begins when it *finishes* -- and once the round has begun it carries the
+    opening stamp, which `is_opening` reads as "still the one that opens the
+    turn". Both are the same answer to this question, and missing the second means
+    a reply that has already been compacted-for and titled gets it again on the
+    next look.
     """
-    return crowd_service.state_of(message) is None
+    return crowd_service.scheduling_state(message) is None
 
 
 def _opens_the_turn_id(generation: Generation) -> bool:

src/lembas/services/prompts.py

@@ -2023,17 +2023,28 @@ BUILTIN: tuple[Fragment, ...] = (
         group=GROUP_TASKS,
         order=451,
         hint="Added as the last turn when a member speaks on the forward pass. "
-        "The failure to word against is a member that repeats what has already "
-        "been said in different words, which is what makes a crowd feel like an "
-        "echo rather than a second opinion.",
+        "Two failures to word against. One is a member that repeats what has "
+        "already been said in different words, which makes a crowd an echo "
+        "rather than a second opinion. The other only shows up on a request that "
+        "asks for something to be *made* -- write this, pick one, draft that -- "
+        "where a member reads the original instruction as addressed to it too "
+        "and produces a rival answer beside its critique. That is not a second "
+        "opinion either; it is two first opinions, and it is what sends a round "
+        "off the question.",
         default=(
             "You are one of several models answering this. The answers above are "
             "quoted with the name of whoever wrote them; yours comes next.\n"
             "\n"
+            "Respond to what is above you. Do not answer the person's original "
+            "request again yourself — that has been done, and your turn is about "
+            "what was done with it.\n"
+            "\n"
             "Add what is missing, correct what is wrong, and say what you would "
-            "have done differently. Do not restate what has already been said to "
-            "show that you agree with it — if you have nothing to add, say so in "
-            "one line and stop. Be brief: somebody is reading all of these."
+            "have done differently and why. Where you would have made a different "
+            "choice, say what it would buy — naming an alternative is not the same "
+            "as giving a reason to prefer it. Do not restate what has already been "
+            "said to show that you agree with it — if you have nothing to add, say "
+            "so in one line and stop. Be brief: somebody is reading all of these."
         ),
     ),
     Fragment(
@@ -2066,11 +2077,23 @@ BUILTIN: tuple[Fragment, ...] = (
         "round. Its own fragment rather than a sentence inside the one below, "
         "because inviting a choice a model cannot express is worse than not "
         "offering it: on a model without the tools capability there is no "
-        "crowd_again to call, and that is the case the next fragment covers.",
+        "crowd_again to call, and that is the case the next fragment covers.\n"
+        "\n"
+        "The failure to word against is capitulation: the model that opened the "
+        "round abandoning its own answer because somebody spoke after it. A "
+        "closing turn told only to synthesise will follow the last speaker, "
+        "which is how a crowd ends up less accurate than the model that started "
+        "it.",
         default=(
             "You opened this and you are closing it. The others have answered and "
             "have had the chance to disagree.\n"
             "\n"
+            "Your own answer is not automatically the worse one for having been "
+            "written first. Change your position where somebody gave you a reason, "
+            "and say what the reason was; agreement with no argument behind it is "
+            "not a reason, and neither is a member having moved on to something "
+            "else.\n"
+            "\n"
             "Write the answer the person actually asked for. Take what the others "
             "got right, say where you disagree with them and why, and name "
             "anything still unresolved rather than papering over it. Attribute "
@@ -2091,11 +2114,18 @@ BUILTIN: tuple[Fragment, ...] = (
         "is reached, or this model has no tools and so cannot ask. It says the "
         "answer has to be final rather than inviting a choice that would be "
         "ignored, which is the difference between a feature and a feature that "
-        "looks like one.",
+        "looks like one. It carries the same guard against capitulation as the "
+        "fragment above, and for the same reason.",
         default=(
             "You opened this and you are closing it, and this is the last turn: "
             "there will be no further round.\n"
             "\n"
+            "Your own answer is not automatically the worse one for having been "
+            "written first. Change your position where somebody gave you a reason, "
+            "and say what the reason was; agreement with no argument behind it is "
+            "not a reason, and neither is a member having moved on to something "
+            "else.\n"
+            "\n"
             "Write the answer the person actually asked for. Take what the others "
             "got right, say where you disagree with them and why, and attribute "
             "what you took from whom. Where the disagreement is unresolved, say so "

src/lembas/web/i18n/sk.py

@@ -1075,13 +1075,23 @@ MESSAGES.update(
             "potom nemôže držať kolo otvorené celé poobedie."
         ),
         "Fold away a short \"I agree\" on the way back": (
-            "Zbaliť krátke „súhlasím“ na cestě späť"
+            "Zbaliť krátke „súhlasím“ na ceste späť"
         ),
-        "Off by default. With it on, each chat's settings panel offers the other models; a chat with none ticked behaves exactly as it always has.": (
-            "Predvolene vypnuté. Po zapnutí panel nastavení každej konverzácie ponúka "
-            "ostatné modely; konverzácia bez zaškrtnutého modelu sa chová presne ako "
-            "vždy."
+        "Off by default. With it on, every chat's composer offers the other models; a chat with none ticked behaves exactly as it always has.": (
+            "Predvolene vypnuté. Po zapnutí ponúka pole na písanie v každej "
+            "konverzácii ostatné modely; konverzácia bez zaškrtnutého modelu sa chová "
+            "presne ako vždy."
         ),
+        "%(models)s models answer each turn, over up to %(rounds)s rounds.": (
+            "Na každý ťah odpovedá %(models)s modelov, a to najviac v %(rounds)s kolách."
+        ),
+        "%(n)s of %(total)s": "%(n)s z %(total)s",
+        "on the way back": "na ceste späť",
+        "closing": "uzatvára",
+        "round %(n)s": "kolo %(n)s",
+        "no rounds left": "už žiadne kolá",
+        "out of time": "vypršal čas",
+        "two endpoints failed": "dva endpointy zlyhali",
         "Check for due work every": "Kontrolovať splatnú prácu každých",
         "How often it looks": "Ako často sa pozerá",
         "Nothing may repeat faster than": "Nič sa nesmie opakovať častejšie než",

src/lembas/web/static/css/admin.css

@@ -260,7 +260,9 @@ a.tabs__tab { text-decoration: none; }
 */
 .field-row {
   display: grid;
-  grid-template-columns: repeat(auto-fit, minmax(9rem, 1fr));
+  /* Both halves of the pair -- see `.grid--2` in app.css. */
+  min-width: 0;
+  grid-template-columns: repeat(auto-fit, minmax(min(100%, 9rem), 1fr));
   gap: var(--sp-3);
 }
 .field-row > .field { margin-bottom: var(--sp-4); }

src/lembas/web/static/css/app.css

@@ -384,8 +384,16 @@ input.visually-hidden[type="checkbox"] {
 
 /* Multi-column form layout, one definition. */
 .grid { display: grid; gap: var(--sp-4); }
-.grid--2 { grid-template-columns: repeat(auto-fit, minmax(14rem, 1fr)); }
-.grid--3 { grid-template-columns: repeat(auto-fit, minmax(9rem, 1fr)); }
+/* `min(100%, …)` on every auto-fit track and `min-width: 0` with it, for the
+   reason `.suggestions` in chat.css sets out at length. The pair is not
+   optional: `min(100%, …)` stops the track demanding more than the box, and
+   `min-width: 0` stops the *box* demanding more than its parent -- a grid or
+   flex item carries `min-width: auto`, which is a min-content floor, and a
+   floor beats `width`. A stylesheet cannot tell whether one of these grids has
+   been dropped into a flex parent today, so both go on every one of them.
+   `tests/test_narrow_grids.py` refuses a track that has only half the pair. */
+.grid--2 { min-width: 0; grid-template-columns: repeat(auto-fit, minmax(min(100%, 14rem), 1fr)); }
+.grid--3 { min-width: 0; grid-template-columns: repeat(auto-fit, minmax(min(100%, 9rem), 1fr)); }
 
 /* --- Alerts --------------------------------------------------------------- */
 .alert {

src/lembas/web/static/css/chat.css

@@ -266,7 +266,27 @@
 */
 .suggestions {
   display: grid;
-  grid-template-columns: repeat(auto-fit, minmax(13rem, 1fr));
+  /* 🚨 `min-width: 0` is what keeps this grid on the screen, and `width: 100%`
+     alone did not: it is a grid item of `.thread__intro`, so it carries
+     `min-width: auto`, which for a grid item means *a min-content floor* -- and
+     min-width beats width. Its min-content size is two cards side by side, so it
+     rendered 428px wide inside a 390px phone with `width: 100%` set and ignored.
+
+     That floor is also why writing the track as `minmax(min(100%, 13rem), 1fr)`
+     -- the tree's standing rule, and right -- made it *worse* on its own, 428px
+     to 455px: a percentage is indefinite while the floor is being measured, so
+     the track fell back to a card's max-content and raised the very number that
+     was overflowing. The two go together. With the floor removed, `width: 100%`
+     finally resolves against the 366px column, `min(100%, …)` hands the track
+     366px to clamp against, and `auto-fit` places one column.
+
+     It scrolled `.thread-scroll` rather than the page, which is why a pass
+     looking for a document that scrolls sideways never saw it: `overflow-y: auto`
+     makes the other axis scrollable too. Reported on a phone, found by asking
+     which *element* could scroll and then reading its computed `width` against
+     its parent's. */
+  min-width: 0;
+  grid-template-columns: repeat(auto-fit, minmax(min(100%, 13rem), 1fr));
   gap: var(--sp-3);
   width: 100%;
   max-width: 40rem;

src/lembas/web/templates/admin/_layout.html

@@ -50,6 +50,13 @@
           {{ icon("sliders", "icon--sm") }}
           <span class="nav-item__label">{{ t("Models") }}</span>
         </a>
+        {# Its own entry rather than a card on Agents, where it started. Sitting
+           there made it read as an agent-chat feature -- which is what the owner
+           took it for, reasonably, since that is what the page is called. #}
+        <a class="nav-item {{ 'is-active' if section == 'crowd' }}" href="/admin/crowd">
+          {{ icon("users", "icon--sm") }}
+          <span class="nav-item__label">{{ t("A crowd") }}</span>
+        </a>
         <a class="nav-item {{ 'is-active' if section == 'audio' }}" href="/admin/audio">
           {{ icon("speaker", "icon--sm") }}
           <span class="nav-item__label">{{ t("Audio") }}</span>

src/lembas/web/templates/admin/agents.html

@@ -395,76 +395,6 @@
   button was pressed, which is what keeps each group's save handler writing one
   key.
 #}
-{# A third settings group on this page, saved by its own form -- the reason the
-   Helpers card gives. A crowd is not an agent-chat feature either, but this is the
-   page somebody opens to find out what one turn may set going. #}
-<form method="post" action="/admin/agents/crowd" class="form-grid">
-  <section class="card">
-    <h2 class="card__title">{{ t("A crowd") }}</h2>
-    <p class="field__hint">
-      A chat can have more than one model in it. The chat's own model answers, then
-      each of the others in turn; then the order runs <strong>{{ t("backwards") }}</strong>,
-      each one asked whether it disagrees with anything; and it ends back at the
-      first, which either closes or sends them round again.
-    </p>
-
-    <div class="alert">
-      {{ icon("warning", "icon--sm") }}
-      <span>
-        One turn costs <strong>{{ t("models × rounds × 2 − 1") }}</strong> replies — four
-        models over two rounds is fifteen — and on a single local endpoint every
-        change of speaker also loads a different model. Larger crowds of smaller
-        models, and sometimes of bigger ones, start going round in circles: that is
-        what the round limit is for, and it is a limit ordinary work will reach
-        rather than a runaway backstop.
-      </span>
-    </div>
-
-    <div class="field">
-      <label class="checkbox">
-        <input type="checkbox" name="enabled" value="true"
-               {{ 'checked' if crowd.enabled }}>
-        <span>{{ t("Let a chat have a crowd") }}</span>
-      </label>
-      <p class="field__hint">{{ t("Off by default. With it on, each chat's settings panel offers the other models; a chat with none ticked behaves exactly as it always has.") }}</p>
-    </div>
-
-    <div class="field">
-      <label class="field__label" for="crowd_max_models">{{ t("Most models besides the chat's own") }}</label>
-      <input class="input" id="crowd_max_models" name="max_models"
-             type="number" min="1" max="8" step="1" value="{{ crowd.max_models }}">
-      <p class="field__hint">{{ t("Four is already eight replies a turn at one round each. More voices past that tend to repeat each other rather than add anything.") }}</p>
-    </div>
-
-    <div class="field">
-      <label class="field__label" for="crowd_max_rounds">{{ t("Most rounds") }}</label>
-      <input class="input" id="crowd_max_rounds" name="max_rounds"
-             type="number" min="1" max="5" step="1" value="{{ crowd.max_rounds }}">
-      <p class="field__hint">{{ t("A round is out and back. Two gives the first model one chance to change its mind after hearing the objections, which is the point of the whole thing; three is where going in circles starts.") }}</p>
-    </div>
-
-    <div class="field">
-      <label class="field__label" for="crowd_wall_seconds">{{ t("Longest a turn may take") }}</label>
-      <input class="input" id="crowd_wall_seconds" name="wall_seconds"
-             type="number" min="60" max="7200" step="30" value="{{ crowd.wall_seconds }}">
-      <p class="field__hint">{{ t("Across every speaker, not each. A member whose endpoint has stalled cannot then hold the round open all afternoon.") }}</p>
-    </div>
-
-    <div class="field">
-      <label class="checkbox">
-        <input type="checkbox" name="collapse_agreement" value="true"
-               {{ 'checked' if crowd.collapse_agreement }}>
-        <span>{{ t('Fold away a short "I agree" on the way back') }}</span>
-      </label>
-      <p class="field__hint">{{ t("The disagreements are what a crowd is for; a column of bubbles saying nothing is what makes somebody switch it off. The text is still there behind a disclosure.") }}</p>
-    </div>
-
-    <div class="btn-row">
-      <button class="btn btn--primary" type="submit">{{ t("Save") }}</button>
-    </div>
-  </section>
-</form>
-
 <form method="post" action="/admin/agents/subagents" class="form-grid">
   <section class="card">
     <h2 class="card__title">{{ t("Helpers") }}</h2>

src/lembas/web/templates/admin/crowd.html

@@ -0,0 +1,87 @@
+{% extends "admin/_layout.html" %}
+{% from "_macros.html" import icon %}
+{% set section = "crowd" %}
+
+{% block title %}A crowd - {{ brand.name }}{% endblock %}
+{% block heading %}A crowd{% endblock %}
+
+{% block admin_content %}
+{#
+  Its own page rather than a card on Agents, which is where it shipped in 1.6.0.
+  Sitting there made it read as an agent-chat feature -- the owner took it for one,
+  reasonably, because that is what the page is called -- and a crowd has nothing to
+  do with agent chats: it works in any conversation.
+#}
+<p class="admin-lede">{{ t("Several models answering one turn, in any chat. Not an agent-chat feature: it works in an ordinary conversation, and the control is in the composer beside the tool switches.") }}</p>
+
+{% if saved %}
+<div class="alert alert--success">{{ icon("check", "icon--sm") }} <span>{{ t("Settings saved.") }}</span></div>
+{% endif %}
+
+<form method="post" action="/admin/crowd" class="form-grid">
+  <section class="card">
+        <p class="field__hint">
+      A chat can have more than one model in it. The chat's own model answers, then
+      each of the others in turn; then the order runs <strong>{{ t("backwards") }}</strong>,
+      each one asked whether it disagrees with anything; and it ends back at the
+      first, which either closes or sends them round again.
+    </p>
+
+    <div class="alert">
+      {{ icon("warning", "icon--sm") }}
+      <span>
+        One turn costs <strong>{{ t("models × rounds × 2 − 1") }}</strong> replies — four
+        models over two rounds is fifteen — and on a single local endpoint every
+        change of speaker also loads a different model. Larger crowds of smaller
+        models, and sometimes of bigger ones, start going round in circles: that is
+        what the round limit is for, and it is a limit ordinary work will reach
+        rather than a runaway backstop.
+      </span>
+    </div>
+
+    <div class="field">
+      <label class="checkbox">
+        <input type="checkbox" name="enabled" value="true"
+               {{ 'checked' if crowd.enabled }}>
+        <span>{{ t("Let a chat have a crowd") }}</span>
+      </label>
+      <p class="field__hint">{{ t("Off by default. With it on, every chat's composer offers the other models; a chat with none ticked behaves exactly as it always has.") }}</p>
+    </div>
+
+    <div class="field">
+      <label class="field__label" for="crowd_max_models">{{ t("Most models besides the chat's own") }}</label>
+      <input class="input" id="crowd_max_models" name="max_models"
+             type="number" min="1" max="8" step="1" value="{{ crowd.max_models }}">
+      <p class="field__hint">{{ t("Four is already eight replies a turn at one round each. More voices past that tend to repeat each other rather than add anything.") }}</p>
+    </div>
+
+    <div class="field">
+      <label class="field__label" for="crowd_max_rounds">{{ t("Most rounds") }}</label>
+      <input class="input" id="crowd_max_rounds" name="max_rounds"
+             type="number" min="1" max="5" step="1" value="{{ crowd.max_rounds }}">
+      <p class="field__hint">{{ t("A round is out and back. Two gives the first model one chance to change its mind after hearing the objections, which is the point of the whole thing; three is where going in circles starts.") }}</p>
+    </div>
+
+    <div class="field">
+      <label class="field__label" for="crowd_wall_seconds">{{ t("Longest a turn may take") }}</label>
+      <input class="input" id="crowd_wall_seconds" name="wall_seconds"
+             type="number" min="60" max="7200" step="30" value="{{ crowd.wall_seconds }}">
+      <p class="field__hint">{{ t("Across every speaker, not each. A member whose endpoint has stalled cannot then hold the round open all afternoon.") }}</p>
+    </div>
+
+    <div class="field">
+      <label class="checkbox">
+        <input type="checkbox" name="collapse_agreement" value="true"
+               {{ 'checked' if crowd.collapse_agreement }}>
+        <span>{{ t('Fold away a short "I agree" on the way back') }}</span>
+      </label>
+      <p class="field__hint">{{ t("The disagreements are what a crowd is for; a column of bubbles saying nothing is what makes somebody switch it off. The text is still there behind a disclosure.") }}</p>
+    </div>
+
+    <div class="btn-row">
+      <button class="btn btn--primary" type="submit">{{ t("Save") }}</button>
+    </div>
+  </section>
+</form>
+
+{% endblock %}

src/lembas/web/templates/chat/_composer.html

@@ -293,6 +293,97 @@
             </div>
           </div>
           {% endif %}
+
+          {#
+            Who else answers.
+
+            Beside the tool switches rather than buried in Chat settings, and
+            *inside this form* rather than in the topbar, for one reason each.
+
+            The first: somebody deciding who answers is making the same kind of
+            choice as somebody picking the model, and the first version of this
+            put it only in the Chat settings panel — behind the ⋯ menu, inside a
+            chat that already existed. The owner enabled the feature, went
+            looking, and could not find it. A control nobody can find is a
+            feature nobody has.
+
+            The second: on the new-chat screen there is no chat row to attach
+            anybody to, so the choice has to *ride along with the first message*
+            — which means being a field of this form. That is the same mechanism
+            the scope switches above use, with the same hidden-input trick,
+            because a browser submits only the ticked boxes and `start_chat`
+            needs to know which ones were not.
+          #}
+          {% if crowd_available %}
+          <div class="picker picker--up" data-picker>
+            <button class="btn btn--icon composer__btn" type="button" data-picker-toggle
+                    aria-haspopup="menu" aria-expanded="false"
+                    aria-label="{{ t('Crowd') }}" title="{{ t('Crowd') }}">
+              {{ icon("users") }}
+              {% if crowd_member_ids %}
+              <span class="composer__count">{{ crowd_member_ids|length + 1 }}</span>
+              {% endif %}
+            </button>
+
+            <div class="picker__menu picker__menu--scope" data-picker-menu role="menu"
+                 hidden aria-label="{{ t('Crowd') }}">
+              <p class="picker__lede">
+                {{ t("Tick a model to have it answer after this one, then be asked whether it disagrees.") }}
+              </p>
+              <p class="picker__group">{{ t("Also answering") }}</p>
+              {% if chat %}
+              {# One hidden field for the whole list, always submitted, so
+                 unticking the last box still says something -- an absent checkbox
+                 carries no signal of its own. #}
+              <input type="hidden" name="crowd_model_ids" value="" form="crowd-form">
+              {% else %}
+              <input type="hidden" name="crowd_model_ids" value="">
+              {% endif %}
+              {% for model in crowd_available %}
+              <label class="picker__option picker__option--toggle">
+                {% if chat %}
+                {# An existing chat: written at once. The verb is on the checkbox
+                   and not on `#crowd-form`, because htmx binds a trigger to the
+                   annotated element and `change` bubbles through *ancestors* --
+                   which a sibling form is not. `form=` scopes the values, and
+                   only the values: without it the PATCH would carry the
+                   composer's own `content` and `project_dir`, and `update_chat`
+                   answers that with a 409. The same reasoning the agent mode
+                   select below carries. #}
+                <input type="checkbox" name="crowd_model_ids" value="{{ model.model_id }}"
+                       {{ 'checked' if model.model_id in crowd_member_ids }}
+                       form="crowd-form"
+                       hx-patch="/api/chats/{{ chat.id }}" hx-swap="none">
+                {% else %}
+                <input type="checkbox" name="crowd_model_ids" value="{{ model.model_id }}"
+                       {{ 'checked' if model.model_id in crowd_member_ids }}>
+                {% endif %}
+                <span class="picker__option-body">
+                  <span class="picker__option-name">{{ model.label }}</span>
+                  {% if model.description %}
+                  <span class="picker__option-note">{{ model.description }}</span>
+                  {% endif %}
+                </span>
+              </label>
+              {% endfor %}
+              {% if crowd_member_ids %}
+              {# One sentence and not three, with the numbers as placeholders: a
+                 translation puts the parts in its own order, and two of these
+                 fragments are not sentences in any language. #}
+              <p class="picker__lede">
+                {{ t("%(models)s models answer each turn, over up to %(rounds)s rounds.",
+                      models=crowd_member_ids|length + 1, rounds=crowd_rounds) }}
+              </p>
+              {% endif %}
+              {% if crowd_skipped %}
+              <p class="picker__lede">
+                {{ t("Skipped, because you cannot reach them any more:") }}
+                <s>{{ crowd_skipped|join(", ") }}</s>
+              </p>
+              {% endif %}
+            </div>
+          </div>
+          {% endif %}
         </div>
 
         {#
@@ -502,6 +593,12 @@
        hx-patch and not hx-post: there is no POST for a chat, only PATCH, and
        htmx shows nothing when a request 405s -- which is how these controls
        spent the first half of their lives doing nothing. #}
+    {% if chat and crowd_available %}
+    {# Empty, and a sibling of the composer's form rather than inside it. See the
+       crowd checkboxes above, and `#agent-mode-form` below, for why both halves
+       of that sentence matter. #}
+    <form id="crowd-form"></form>
+    {% endif %}
     {% if chat and chat.kind == "agent" %}
     <form id="agent-mode-form"></form>
     {% endif %}

src/lembas/web/templates/chat/_message.html

@@ -88,24 +88,24 @@
          replies: a round produces more bubbles than it has models in it. #}
       <span class="badge">
         {% if crowd.get("phase") == "out" %}
-          {{ crowd.get("index", 0) + 1 }} of {{ crowd.get("of", 1) }}
+          {{ t("%(n)s of %(total)s", n=crowd.get("index", 0) + 1, total=crowd.get("of", 1)) }}
         {% elif crowd.get("phase") == "back" %}
-          on the way back
+          {{ t("on the way back") }}
         {% else %}
-          closing
+          {{ t("closing") }}
         {% endif %}
-        {% if crowd.get("round", 1) > 1 %} · round {{ crowd.get("round") }}{% endif %}
+        {% if crowd.get("round", 1) > 1 %} · {{ t("round %(n)s", n=crowd.get("round")) }}{% endif %}
       </span>
       {% if crowd.get("stopped") %}
       {# Why a round ended, where it ended. Without this a crowd that ran out of
          rounds or time simply stops, which reads as the feature failing. #}
       <span class="badge badge--warning" title="{{ t('The round ended here') }}">
         {% if crowd.get("stopped") == "rounds" %}
-          no rounds left
+          {{ t("no rounds left") }}
         {% elif crowd.get("stopped") == "time" %}
-          out of time
+          {{ t("out of time") }}
         {% else %}
-          two endpoints failed
+          {{ t("two endpoints failed") }}
         {% endif %}
       </span>
       {% endif %}

tests/test_crowd_chain.py

@@ -177,6 +177,73 @@ def test_the_round_is_recorded_on_every_row(db, started):
     assert len(anchors) == 1
 
 
+def test_the_reply_that_opened_the_round_is_stamped_too(db, started):
+    """The opening bubble says `1 of 3` like every other one.
+
+    It is the one contribution the crowd does not start -- the composer does --
+    so until the round begins there is nothing to stamp it with. Before this, a
+    two-model round rendered as an unmarked reply followed by one saying `2 of 2`,
+    with no 1 anywhere.
+    """
+    chat = _crowd_chat(db)
+    opening = _opening_reply(db, chat)
+    assert crowd_service.state_of(opening) is None, "nothing to say before it finishes"
+
+    assert _advance(db, chat, opening)
+    db.expire_all()
+
+    state = crowd_service.state_of(opening)
+    assert state is not None
+    assert (state.phase, state.index) == (crowd_service.PHASE_OUT, 0)
+    assert state.of == 3
+
+
+def test_the_opening_stamp_belongs_to_the_same_round(db, started):
+    chat = _crowd_chat(db)
+    opening = _opening_reply(db, chat)
+    order = []
+    assert _advance(db, chat, opening)
+    db.expire_all()
+    order = _incomplete(db, chat)
+
+    opened = crowd_service.state_of(opening)
+    first = crowd_service.state_of(order[0])
+    # Same question, same clock -- or the chips group two bubbles of one round
+    # under two different rounds.
+    assert opened.turn == first.turn
+    assert opened.started_at == first.started_at
+    assert opened.round == first.round == 1
+
+
+def test_the_opening_stamp_is_not_scheduling_state(db, started):
+    """It must read as "no round yet" everywhere that decides what happens next.
+
+    Fed to the scheduler it would be a member at index 0, which inherits the old
+    `started_at` -- so regenerating the opening an hour later would end the round
+    with "out of time" before anybody spoke -- and it would hand that reply a
+    member's tools and a member's instruction instead of an ordinary first answer.
+    """
+    chat = _crowd_chat(db)
+    opening = _opening_reply(db, chat)
+    assert _advance(db, chat, opening)
+    db.expire_all()
+
+    assert crowd_service.state_of(opening) is not None
+    assert crowd_service.scheduling_state(opening) is None
+    assert crowd_service.is_opening(crowd_service.state_of(opening))
+    assert generation_service._opens_the_turn(opening)
+
+
+def test_a_later_speaker_is_not_mistaken_for_the_opening(db, started):
+    chat = _crowd_chat(db)
+    order = _run_round(db, chat, started)
+    for message in order:
+        state = crowd_service.state_of(message)
+        assert not crowd_service.is_opening(state)
+        # `==` and not `is`: `state_of` builds a fresh Turn on every call.
+        assert crowd_service.scheduling_state(message) == state
+
+
 def test_each_speaker_carries_its_own_connection(db, started):
     """So `speaker_for` resolves the pair rather than guessing at the id."""
     chat = _crowd_chat(db)

tests/test_crowd_ui.py

@@ -70,6 +70,87 @@ def _members(db, chat) -> list[str]:
     ]
 
 
+# --- Reachable where somebody would look --------------------------------------
+#
+# The feature shipped in 1.6.0 switched on and unreachable: the only control was
+# inside the Chat settings panel, behind the ⋯ menu, in a chat that already
+# existed. The owner enabled it, went looking, and reported that there was nothing
+# to find. A control nobody can find is a feature nobody has, so these assert the
+# two places it has to be rather than the one place it was.
+def test_the_composer_offers_the_crowd_in_a_chat(client, db):
+    """Beside the tool switches, where the comparable decisions are."""
+    chat = _chat(db)
+    page = client.get(f"/chat/{chat.id}").text
+    assert 'name="crowd_model_ids"' in page
+    assert 'form="crowd-form"' in page
+    assert '<form id="crowd-form">' in page
+
+
+def test_the_composer_offers_the_crowd_before_the_chat_exists(client, db):
+    """On the new-chat screen there is no row to attach anybody to, so the choice
+    rides along with the first message — the mechanism the scope switches use."""
+    page = client.get("/chat").text
+    assert 'name="crowd_model_ids"' in page
+    # Riding along, so no sibling form and no PATCH: the composer's own POST
+    # carries it.
+    assert '<form id="crowd-form">' not in page
+    assert 'value="second-model"' in page
+
+
+def test_starting_a_chat_with_a_crowd_keeps_it(client, db):
+    """The end of that path: the first message creates the chat *and* its crowd."""
+    from sqlalchemy import select as sa_select
+
+    client.post(
+        "/api/chats/start",
+        data={
+            "content": "Who is right?",
+            "model_id": "main-model",
+            "crowd_model_ids": ["", "second-model", "third-model"],
+        },
+        follow_redirects=False,
+    )
+    db.expire_all()
+    chat = db.scalars(sa_select(Chat).order_by(Chat.created_at.desc())).first()
+    assert [row.model_id for row in sorted(chat.crowd, key=lambda r: r.position)] == [
+        "second-model",
+        "third-model",
+    ]
+
+
+def test_starting_a_chat_refuses_a_model_the_person_cannot_reach(client, db):
+    """The same rule as the panel, in the same one function, so there is nowhere
+    for the two to disagree."""
+    from sqlalchemy import select as sa_select
+
+    group = Group(name="Wheel")
+    db.add(group)
+    restricted = db.scalar(select(Model).where(Model.model_id == "third-model"))
+    restricted.public = False
+    restricted.groups = [group]
+    user = _user(db)
+    user.role = "user"
+    db.commit()
+
+    client.post(
+        "/api/chats/start",
+        data={
+            "content": "Who is right?",
+            "model_id": "main-model",
+            "crowd_model_ids": ["second-model", "third-model"],
+        },
+        follow_redirects=False,
+    )
+    db.expire_all()
+    chat = db.scalars(sa_select(Chat).order_by(Chat.created_at.desc())).first()
+    assert [row.model_id for row in chat.crowd] == ["second-model"]
+
+
+def test_the_composer_control_is_absent_while_the_feature_is_off(client, db):
+    settings_store.update(db, {"enabled": False}, key=settings_store.CROWD)
+    assert 'name="crowd_model_ids"' not in client.get("/chat").text
+
+
 # --- Choosing -----------------------------------------------------------------
 def test_the_panel_offers_the_other_models(client, db):
     chat = _chat(db)
@@ -200,6 +281,31 @@ def test_a_bubble_on_the_way_out_says_which_speaker_it_is(db):
     assert "2 of 3" in html
 
 
+def test_the_bubble_that_opened_the_round_says_it_is_first(db):
+    """The opening reply is stamped once the round begins, so it says `1 of 3`.
+
+    Before that it was the one contribution with no chip at all, which made a
+    two-model round read as an ordinary answer followed by one labelled `2 of 2`.
+    """
+    chat = _chat(db)
+    html = _bubble(db, chat, phase=crowd_service.PHASE_OUT, index=0, of=3)
+    assert "1 of 3" in html
+
+
+def test_the_chip_is_translated(db):
+    """It is prose a person reads, and it was English on a Slovak instance."""
+    from lembas.web import i18n
+
+    chat = _chat(db)
+    i18n.activate("sk")
+    try:
+        html = _bubble(db, chat, phase=crowd_service.PHASE_BACK, index=1, of=3)
+    finally:
+        i18n.activate("en")
+    assert "na ceste späť" in html
+    assert "on the way back" not in html
+
+
 def test_a_bubble_on_the_way_back_says_so_and_is_quieter(db):
     chat = _chat(db)
     html = _bubble(db, chat, phase=crowd_service.PHASE_BACK)

tests/test_narrow_grids.py

@@ -0,0 +1,85 @@
+"""A grid that reflows can still overflow a phone, and twice it has.
+
+`repeat(auto-fit, minmax(13rem, 1fr))` puts two 208px cards side by side on a
+390px screen: `auto-fit` decides how many columns fit, and a track whose minimum
+is a fixed length never gives that minimum up. The tree's standing rule is to
+write `minmax(min(100%, 13rem), 1fr)` instead, so the *track* yields rather than
+the viewport.
+
+Half of the rule is not the rule. `.suggestions` had `width: 100%` and got the
+`min(100%, …)` track, and still rendered 455px wide inside a 366px column -- it
+is a grid item, so it carries `min-width: auto`, which for a grid item means a
+min-content floor, and a floor beats `width`. Worse, the floor is measured while
+the percentage is indefinite, so the track falls back to a card's max-content:
+adding `min(100%, …)` on its own moved the overflow from 428px to 455px.
+
+So both halves are asserted here, on every auto-fit grid in the stylesheets. A
+stylesheet cannot see whether a given grid is a flex item on some page today or
+becomes one next week, and `min-width: 0` costs nothing where it is not needed.
+
+Found on a phone, on the one screen the screenshot harness had never actually
+rendered -- `scripts/shoot.py` builds its client without a lifespan, so the
+suggestion cards were absent from every shot ever taken of the new-chat screen.
+That is fixed there; this file is the cheap half that runs in the suite.
+"""
+
+from __future__ import annotations
+
+import re
+from pathlib import Path
+
+import lembas
+
+CSS = Path(lembas.__file__).parent / "web/static/css"
+
+# `[^{}]*` cannot cross a brace, so an `@media` prelude never matches and the
+# rules nested inside it do.
+RULE = re.compile(r"([^{}]*)\{([^{}]*)\}")
+COMMENT = re.compile(r"/\*.*?\*/", re.S)
+
+
+def _auto_grids() -> list[tuple[str, str, dict[str, str]]]:
+    found = []
+    for path in sorted(CSS.glob("*.css")):
+        text = COMMENT.sub("", path.read_text(encoding="utf-8"))
+        for prelude, body in RULE.findall(text):
+            if "auto-fit" not in body and "auto-fill" not in body:
+                continue
+            declarations = {
+                part.partition(":")[0].strip(): part.partition(":")[2].strip()
+                for part in body.split(";")
+                if ":" in part
+            }
+            found.append((path.name, prelude.strip(), declarations))
+    return found
+
+
+def test_the_stylesheets_still_have_auto_fit_grids_to_check():
+    # Otherwise the two tests below pass by finding nothing, which is how a
+    # coverage test quietly stops covering anything.
+    assert len(_auto_grids()) >= 4
+
+
+def test_every_auto_fit_track_can_give_up_its_minimum():
+    offenders = [
+        f"{name} {selector}"
+        for name, selector, declarations in _auto_grids()
+        for value in [declarations.get("grid-template-columns", "")]
+        if "minmax(" in value and "min(100%" not in value
+    ]
+    assert not offenders, (
+        "an auto-fit track with a fixed minimum overflows a phone; write "
+        f"minmax(min(100%, X), 1fr): {offenders}"
+    )
+
+
+def test_every_auto_fit_grid_drops_its_automatic_minimum_size():
+    offenders = [
+        f"{name} {selector}"
+        for name, selector, declarations in _auto_grids()
+        if declarations.get("min-width") != "0"
+    ]
+    assert not offenders, (
+        "a grid item's `min-width: auto` is a min-content floor and beats "
+        f"`width`, so `min(100%, …)` alone does not save it: {offenders}"
+    )