enderofwings/NexusOS
1
0
Fork 1
Code Issues Pull Requests 1 Actions 10 Packages Projects Releases Wiki Activity

feat(models): gate optional models behind required mistral+nomic-embed

Browse Source 
Models page redesign: tabs (Required/Recommended/Installed) with a
multi-column card grid, required models (memory curator + embeddings)
surfaced first and gating the rest until both are installed. Adds a
qwen2.5:14b tier to the hardware-fit catalog for high-VRAM machines.
Installer and restore messaging updated to match. Also declares pytest
in requirements-base.txt so bin/check.sh's test suite is reproducible
on a fresh venv.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Jon Wingender
2026-07-27 19:45:30 -05:00
co-authored by Claude Sonnet 5
parent 7d9681907c
commit 63c93346ae
5 changed files with 355 additions and 160 deletions
Download Patch File Download Diff File Expand all files Collapse all files
bin/sync.py
+3 -1
View File
@@ -241,7 +241,9 @@ def cmd_restore(args) -> int:
if not args.no_desktop:
linux_stage("restore-linux.sh", "desktop")
print("\nRestore complete. Nexus is ready to start.")
print("Note: Ollama models are not in the backup - pull them with `ollama pull <model>`.")
print("Note: Ollama models are not in the backup. Open the Models tab -> Required")
print("and pull the two models NexusOS depends on (memory curator + embeddings)")
print("before anything else, then pick a hardware-fit chat model.")
return 0
install-windows.ps1
+7 -6
View File
@@ -321,12 +321,13 @@ else { Write-Warn "Could not persist default model - pick it at the top of the c
# an 8B model). The Models tab detects VRAM/RAM and marks which models fit, so
# the user pulls the right ones there instead of us guessing several GB.
Write-Step "Skipping model download (pick hardware-appropriate models in the app)"
Write-Host " No models were downloaded. Open NexusOS -> Models: it detects your" -ForegroundColor DarkGray
Write-Host " VRAM/RAM and flags which models fit (green = GPU, yellow = CPU/RAM)." -ForegroundColor DarkGray
Write-Host " Pull at least:" -ForegroundColor DarkGray
Write-Host " - $EmbedModel (required for recall / document search)" -ForegroundColor Gray
Write-Host " - a chat model the Models tab marks as fitting your GPU (or $ChatModel on a big one)" -ForegroundColor Gray
Write-Host " - $MemModel for the memory curator (optional)" -ForegroundColor Gray
Write-Host " No models were downloaded. Open NexusOS -> Models. The Required tab" -ForegroundColor DarkGray
Write-Host " lists the two models NexusOS itself depends on and gates the rest" -ForegroundColor DarkGray
Write-Host " of the catalog until both are installed:" -ForegroundColor DarkGray
Write-Host " - $MemModel (required - memory curator: extracts facts, titles chats)" -ForegroundColor Gray
Write-Host " - $EmbedModel (required - semantic recall of past conversations)" -ForegroundColor Gray
Write-Host " After that, the Recommended tab detects your VRAM/RAM and flags which" -ForegroundColor DarkGray
Write-Host " chat models fit (green = GPU, yellow = CPU/RAM), e.g. $ChatModel." -ForegroundColor DarkGray
# -- Make Ollama manual-start (NexusOS owns the lifecycle) ----------------------
Write-Step "Setting Ollama to manual start"
interface/web/src/Models.jsx
+332 -150
View File
@@ -1,14 +1,101 @@
import { useCallback, useEffect, useState } from "react";
import { useCallback, useEffect, useMemo, useState } from "react";
import { API_BASE } from "./config";
const FIT = {
gpu: { label: "🟢 fits GPU", color: "#8aff8a" },
ram: { label: "🟡 runs on RAM (CPU)", color: "#e8c65a" },
no: { label: "🔴 too big", color: "#ff8a80" },
};
const TABS = [
{ key: "required", label: "Required" },
{ key: "recommended", label: "Recommended" },
{ key: "installed", label: "Installed" },
];
const GRID_STYLE = {
display: "grid",
gridTemplateColumns: "repeat(auto-fill, minmax(260px, 1fr))",
gap: "0.75rem",
alignItems: "stretch",
};
function ModelCard({ model, installed, pulling, pullingName, locked, onPull }) {
const fit = FIT[model.fit] || FIT.no;
const isPullingThis = pulling && pullingName === model.name;
return (
<div
style={{
display: "flex",
flexDirection: "column",
gap: "0.5rem",
padding: "0.85rem",
background: locked ? "#131313" : "#0f0f0f",
border: `1px solid ${model.required ? "#3a5a3a" : "#262626"}`,
borderRadius: "10px",
opacity: locked ? 0.45 : 1,
transition: "opacity 0.15s ease",
}}
>
<div style={{ display: "flex", justifyContent: "space-between", alignItems: "flex-start", gap: "0.5rem" }}>
<div style={{ minWidth: 0 }}>
<div style={{ color: "#eee", fontSize: "0.9rem", fontWeight: 500, wordBreak: "break-word" }}>
{model.name}
</div>
<div style={{ color: "#666", fontSize: "0.75rem" }}>
{model.params} · {model.size_gb} GB · {model.role}
</div>
</div>
{model.required && (
<span style={{
fontSize: "0.65rem", color: "#8aff8a", border: "1px solid #3a5a3a",
borderRadius: "999px", padding: "0.1rem 0.5rem", whiteSpace: "nowrap",
}}>
required
</span>
)}
</div>
<div style={{ color: "#888", fontSize: "0.75rem", flexGrow: 1 }}>{model.note}</div>
<div style={{ display: "flex", justifyContent: "space-between", alignItems: "center", gap: "0.5rem" }}>
<span style={{ color: fit.color, fontSize: "0.72rem", whiteSpace: "nowrap" }}>{fit.label}</span>
{installed ? (
<span style={{ color: "#8aff8a", fontSize: "0.75rem", whiteSpace: "nowrap" }}>✓ installed</span>
) : (
<button
onClick={() => onPull(model.name)}
disabled={pulling || locked}
title={locked ? "Install the required models first" : undefined}
style={{
padding: "0.3rem 0.7rem",
background: pulling || locked ? "#555" : "#28a745",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: pulling || locked ? "not-allowed" : "pointer",
fontSize: "0.75rem",
whiteSpace: "nowrap",
}}
>
{isPullingThis ? "Pulling..." : "Pull"}
</button>
)}
</div>
</div>
);
}
export function Models({ onPullStateChange }) {
const [models, setModels] = useState([]);
const [models, setModels] = useState([]); // installed, chat-facing (embed models hidden)
const [allInstalled, setAllInstalled] = useState([]); // installed, unfiltered (includes embed)
const [loading, setLoading] = useState(false);
const [error, setError] = useState("");
const [pulling, setPulling] = useState(false);
const [pullingName, setPullingName] = useState("");
const [modelName, setModelName] = useState("");
const [pullProgress, setPullProgress] = useState("");
const [recommended, setRecommended] = useState(null); // {hardware, models}
const [tab, setTab] = useState("required");
const [manualOpen, setManualOpen] = useState(false);
const checkOllamaStatus = useCallback(async () => {
try {
@@ -28,16 +115,21 @@ export function Models({ onPullStateChange }) {
if (!isRunning) {
setError("Ollama service is not running");
setModels([]);
setAllInstalled([]);
return;
}
const response = await fetch(`${API_BASE}/models/details`);
if (!response.ok) throw new Error("Failed to fetch models");
const data = await response.json();
// Embedding models (nomic-embed-text) are infrastructure — hide them.
setModels((data.models || []).filter((m) => !m.name.toLowerCase().includes("embed")));
const all = data.models || [];
setAllInstalled(all);
// Embedding models (nomic-embed-text) are infrastructure — hide from the
// Installed list, but they still count via allInstalled for the Required tab.
setModels(all.filter((m) => !m.name.toLowerCase().includes("embed")));
} catch (err) {
setError(`Error loading models: ${err.message}`);
setModels([]);
setAllInstalled([]);
} finally {
setLoading(false);
}
@@ -51,6 +143,7 @@ export function Models({ onPullStateChange }) {
}
setPulling(true);
setPullingName(name);
onPullStateChange?.(true);
setError("");
setPullProgress(`Pulling ${name}…`);
@@ -100,6 +193,7 @@ export function Models({ onPullStateChange }) {
setError(`Error pulling model: ${err.message}`);
} finally {
setPulling(false);
setPullingName("");
onPullStateChange?.(false);
}
};
@@ -130,12 +224,21 @@ export function Models({ onPullStateChange }) {
return () => clearInterval(interval);
}, [loadModels]);
const installedNames = new Set(models.map(m => m.name.toLowerCase()));
const FIT = {
gpu: { label: "🟢 fits GPU", color: "#8aff8a" },
ram: { label: "🟡 runs on RAM (CPU)", color: "#e8c65a" },
no: { label: "🔴 too big", color: "#ff8a80" },
};
const installedNames = useMemo(
() => new Set(allInstalled.map(m => m.name.toLowerCase())),
[allInstalled]
);
const requiredModels = useMemo(
() => (recommended?.models ?? []).filter(m => m.required),
[recommended]
);
const otherModels = useMemo(
() => (recommended?.models ?? []).filter(m => !m.required),
[recommended]
);
const requiredMissing = requiredModels.filter(m => !installedNames.has(m.name.toLowerCase()));
const requiredSatisfied = recommended != null && requiredMissing.length === 0;
return (
<div style={{ display: "flex", flexDirection: "column", height: "100%", flexGrow: 1 }}>
@@ -161,159 +264,238 @@ export function Models({ onPullStateChange }) {
padding: "1rem",
display: "flex",
flexDirection: "column",
overflowY: "auto",
overflowX: "hidden"
minHeight: 0,
}}>
{/* Pull Model Section */}
<div style={{ marginBottom: "1.5rem", paddingBottom: "1rem", borderBottom: "1px solid #333" }}>
<h3 style={{ margin: "0 0 0.75rem 0", fontSize: "0.95rem", color: "#bbb" }}>Pull Model from Registry</h3>
<div style={{ display: "flex", gap: "0.5rem" }}>
<input
type="text"
placeholder="e.g., llama3, mistral, gemma"
value={modelName}
onChange={(e) => setModelName(e.target.value)}
onKeyPress={(e) => e.key === "Enter" && !pulling && pullModel()}
disabled={pulling}
style={{
flex: 1,
padding: "0.6rem 0.75rem",
background: "#0a0a0a",
color: "#eee",
border: "1px solid #333",
borderRadius: "6px",
fontSize: "0.9rem"
}}
/>
<button
onClick={pullModel}
disabled={pulling}
style={{
padding: "0.6rem 1.25rem",
background: pulling ? "#666" : "#28a745",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: pulling ? "not-allowed" : "pointer",
fontSize: "0.9rem",
whiteSpace: "nowrap"
}}
>
{pulling ? "Pulling..." : "Pull"}
</button>
<button
onClick={loadModels}
disabled={loading}
style={{
padding: "0.6rem 1.25rem",
background: "#007acc",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: loading ? "not-allowed" : "pointer",
opacity: loading ? 0.6 : 1,
fontSize: "0.9rem",
whiteSpace: "nowrap"
}}
>
{loading ? "Refreshing..." : "Refresh"}
</button>
{/* Header: hardware summary + tabs */}
<div style={{ display: "flex", justifyContent: "space-between", alignItems: "center", flexWrap: "wrap", gap: "0.5rem", marginBottom: "0.85rem" }}>
<div style={{ display: "flex", gap: "0.35rem" }}>
{TABS.map(t => {
const active = tab === t.key;
const badge = t.key === "required" && recommended
? (requiredSatisfied ? "✓" : requiredMissing.length)
: t.key === "installed" ? models.length : null;
return (
<button
key={t.key}
onClick={() => setTab(t.key)}
style={{
padding: "0.45rem 0.9rem",
background: active ? "#28a745" : "transparent",
color: active ? "#fff" : "#aaa",
border: `1px solid ${active ? "#28a745" : "#333"}`,
borderRadius: "999px",
cursor: "pointer",
fontSize: "0.82rem",
display: "flex",
alignItems: "center",
gap: "0.4rem",
}}
>
{t.label}
{badge !== null && badge !== undefined && (
<span style={{
fontSize: "0.68rem",
color: active ? "#fff" : (t.key === "required" && !requiredSatisfied ? "#ff8a80" : "#888"),
}}>
{badge}
</span>
)}
</button>
);
})}
</div>
{pullProgress && (
<div style={{ marginTop: "0.75rem", fontSize: "0.85rem", color: "#999" }}>
{pullProgress}
</div>
)}
</div>
{/* Recommended for your hardware */}
{recommended && (
<div style={{ marginBottom: "1.5rem", paddingBottom: "1rem", borderBottom: "1px solid #333" }}>
<h3 style={{ margin: "0 0 0.25rem 0", fontSize: "0.95rem", color: "#bbb" }}>Recommended for your hardware</h3>
<div style={{ fontSize: "0.78rem", color: "#777", marginBottom: "0.75rem" }}>
{recommended && (
<div style={{ fontSize: "0.78rem", color: "#777" }}>
{recommended.hardware.ram_gb ? `${recommended.hardware.ram_gb} GB RAM` : "RAM unknown"}
{" · "}
{recommended.hardware.vram_gb ? `${recommended.hardware.vram_gb} GB GPU` : "GPU VRAM unknown — sized by RAM"}
</div>
<div style={{ display: "flex", flexDirection: "column", gap: "0.35rem" }}>
{recommended.models.map((m) => {
const installed = installedNames.has(m.name.toLowerCase());
const fit = FIT[m.fit] || FIT.no;
return (
<div key={m.name} style={{ display: "flex", alignItems: "center", gap: "0.6rem", padding: "0.4rem 0.6rem", background: "#0f0f0f", border: "1px solid #262626", borderRadius: "8px" }}>
<div style={{ flex: 1, minWidth: 0 }}>
<span style={{ color: "#eee", fontSize: "0.85rem" }}>{m.name}</span>
<span style={{ color: "#666", fontSize: "0.75rem" }}> · {m.params} · {m.size_gb} GB · {m.role}</span>
<div style={{ color: "#777", fontSize: "0.72rem" }}>{m.note}</div>
</div>
<span style={{ color: fit.color, fontSize: "0.75rem", whiteSpace: "nowrap" }}>{fit.label}</span>
{installed
? <span style={{ color: "#8aff8a", fontSize: "0.75rem", whiteSpace: "nowrap" }}>✓ installed</span>
: <button onClick={() => pullModel(m.name)} disabled={pulling}
style={{ padding: "0.3rem 0.7rem", background: pulling ? "#555" : "#28a745", color: "#fff", border: "none", borderRadius: "6px", cursor: pulling ? "not-allowed" : "pointer", fontSize: "0.75rem", whiteSpace: "nowrap" }}>
Pull
</button>}
</div>
);
})}
</div>
)}
<button
onClick={loadModels}
disabled={loading}
style={{
padding: "0.45rem 0.9rem",
background: "#007acc",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: loading ? "not-allowed" : "pointer",
opacity: loading ? 0.6 : 1,
fontSize: "0.82rem",
}}
>
{loading ? "Refreshing..." : "Refresh"}
</button>
</div>
{pullProgress && (
<div style={{ marginBottom: "0.85rem", fontSize: "0.85rem", color: "#999" }}>
{pullProgress}
</div>
)}
{/* Local Models List */}
<div style={{ paddingRight: "0.5rem" }}>
<h3 style={{ margin: "0 0 1rem 0", fontSize: "0.95rem", color: "#bbb" }}>
Available Models ({models.length})
</h3>
{models.length === 0 ? (
<div style={{ padding: "2rem", textAlign: "center", color: "#666" }}>
{loading ? "Loading models..." : "No models available. Pull a model to get started."}
</div>
{/* Scrollable tab content */}
<div style={{ flexGrow: 1, overflowY: "auto", overflowX: "hidden", paddingRight: "0.25rem" }}>
{!recommended ? (
<div style={{ padding: "2rem", textAlign: "center", color: "#666" }}>Loading catalog...</div>
) : tab === "required" ? (
<>
<div style={{ marginBottom: "0.85rem", fontSize: "0.85rem", color: requiredSatisfied ? "#8aff8a" : "#e8c65a" }}>
{requiredSatisfied
? "All required models are installed. Everything else is unlocked on the Recommended tab."
: "NexusOS needs both of these to fully work (memory + recall). Install them before pulling anything else."}
</div>
<div style={GRID_STYLE}>
{requiredModels.map(m => (
<ModelCard
key={m.name}
model={m}
installed={installedNames.has(m.name.toLowerCase())}
pulling={pulling}
pullingName={pullingName}
locked={false}
onPull={pullModel}
/>
))}
</div>
</>
) : tab === "recommended" ? (
<>
{!requiredSatisfied && (
<div style={{ marginBottom: "0.85rem", fontSize: "0.85rem", color: "#e8c65a" }}>
Dimmed until the Required tab is complete — install those first.
</div>
)}
<div style={GRID_STYLE}>
{otherModels.map(m => (
<ModelCard
key={m.name}
model={m}
installed={installedNames.has(m.name.toLowerCase())}
pulling={pulling}
pullingName={pullingName}
locked={!requiredSatisfied}
onPull={pullModel}
/>
))}
</div>
</>
) : (
<div style={{ display: "grid", gap: "0.75rem" }}>
{models.map((model) => (
<div
key={model.name}
<>
{models.length === 0 ? (
<div style={{ padding: "2rem", textAlign: "center", color: "#666" }}>
{loading ? "Loading models..." : "No models installed yet. Start with the Required tab."}
</div>
) : (
<div style={GRID_STYLE}>
{models.map((model) => (
<div
key={model.name}
style={{
display: "flex",
flexDirection: "column",
gap: "0.5rem",
padding: "0.85rem",
background: "#1b1b1b",
borderRadius: "10px",
border: "1px solid #2a2a2a",
}}
>
<div style={{ fontWeight: 500, color: "#ddd", wordBreak: "break-word" }}>{model.name}</div>
<div style={{ fontSize: "0.8rem", color: "#aaa" }}>
{(model.size / (1024 * 1024 * 1024)).toFixed(2)} GB
</div>
{model.modified_at && (
<div style={{ fontSize: "0.78rem", color: "#888" }}>
Modified: {new Date(model.modified_at).toLocaleDateString()}
</div>
)}
<button
onClick={() => deleteModel(model.name)}
style={{
alignSelf: "flex-start",
padding: "0.4rem 0.9rem",
background: "#d63031",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: "pointer",
fontSize: "0.8rem",
}}
>
Delete
</button>
</div>
))}
</div>
)}
</>
)}
</div>
{/* Manual pull — collapsed by default */}
<div style={{ marginTop: "0.85rem", paddingTop: "0.75rem", borderTop: "1px solid #333" }}>
{!manualOpen ? (
<button
onClick={() => setManualOpen(true)}
style={{
background: "none",
border: "none",
color: "#777",
fontSize: "0.82rem",
cursor: "pointer",
padding: 0,
}}
>
Pull a different model...
</button>
) : (
<div>
<div style={{ display: "flex", justifyContent: "space-between", alignItems: "center", marginBottom: "0.5rem" }}>
<h3 style={{ margin: 0, fontSize: "0.85rem", color: "#bbb" }}>Pull model from registry</h3>
<button
onClick={() => setManualOpen(false)}
style={{ background: "none", border: "none", color: "#666", fontSize: "0.78rem", cursor: "pointer" }}
>
Collapse
</button>
</div>
<div style={{ display: "flex", gap: "0.5rem" }}>
<input
type="text"
placeholder="e.g., llama3, mistral, gemma"
value={modelName}
onChange={(e) => setModelName(e.target.value)}
onKeyPress={(e) => e.key === "Enter" && !pulling && pullModel()}
disabled={pulling}
style={{
padding: "0.9rem",
background: "#1b1b1b",
borderRadius: "8px",
border: "1px solid #2a2a2a",
display: "flex",
justifyContent: "space-between",
alignItems: "center"
flex: 1,
padding: "0.6rem 0.75rem",
background: "#0a0a0a",
color: "#eee",
border: "1px solid #333",
borderRadius: "6px",
fontSize: "0.9rem"
}}
/>
<button
onClick={() => pullModel()}
disabled={pulling}
style={{
padding: "0.6rem 1.25rem",
background: pulling ? "#666" : "#28a745",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: pulling ? "not-allowed" : "pointer",
fontSize: "0.9rem",
whiteSpace: "nowrap"
}}
>
<div>
<div style={{ fontWeight: "500", marginBottom: "0.3rem", color: "#ddd" }}>
{model.name}
</div>
<div style={{ fontSize: "0.85rem", color: "#aaa" }}>
{(model.size / (1024 * 1024 * 1024)).toFixed(2)} GB
</div>
{model.modified_at && (
<div style={{ fontSize: "0.85rem", color: "#888" }}>
Modified: {new Date(model.modified_at).toLocaleDateString()}
</div>
)}
</div>
<button
onClick={() => deleteModel(model.name)}
style={{
padding: "0.5rem 1rem",
background: "#d63031",
color: "#fff",
border: "none",
borderRadius: "6px",
cursor: "pointer",
fontSize: "0.85rem",
whiteSpace: "nowrap"
}}
>
Delete
</button>
</div>
))}
{pulling ? "Pulling..." : "Pull"}
</button>
</div>
</div>
)}
</div>
requirements-base.txt
+5
View File
@@ -45,6 +45,8 @@ python-docx
sqlite-vec
# Local speech-to-text: CTranslate2-based, no torch, keeps dictation on-device.
faster-whisper
# Email client: IMAP read (SMTP send is stdlib). Pure-Python, no native deps.
imap-tools
# Documentation Support
markdown-it-py
@@ -52,3 +54,6 @@ MarkupSafe
mdurl
Pygments
Jinja2
# Testing (bin/check.sh is the release gate; it shells out to pytest)
pytest
synapse/hardware.py
+8 -3
View File
@@ -11,16 +11,21 @@ import subprocess
from typing import Any, Dict, List, Optional
# Curated local-friendly models with their Q4 on-disk sizes (GB) and role.
# Sizes are approximate default-quant download sizes.
# Sizes are approximate default-quant download sizes. `required: True` marks
# the two models NexusOS itself depends on (memory curator + embeddings) —
# see DEFAULT_MEMORY_MODEL / DEFAULT_EMBED_MODEL in nexus_config.py, the
# single source of truth these two entries must stay in sync with. The
# Models page surfaces required models first and gates the rest behind them.
CATALOG: List[Dict[str, Any]] = [
{"name": "mistral:latest", "size_gb": 4.1, "params": "7B", "role": "memory", "note": "Memory curator: extracts facts, titles conversations", "required": True},
{"name": "nomic-embed-text","size_gb": 0.27,"params": "—", "role": "embeddings", "note": "Powers semantic recall of past conversations", "required": True},
{"name": "gemma2:2b", "size_gb": 1.6, "params": "2B", "role": "chat", "note": "Smallest; fast on any GPU"},
{"name": "llama3.2:3b", "size_gb": 2.0, "params": "3B", "role": "chat", "note": "Small Llama, fits 4GB GPU"},
{"name": "qwen3:4b", "size_gb": 2.5, "params": "4B", "role": "chat", "note": "Reasoning (Think toggle); great on a 4GB card"},
{"name": "phi3:mini", "size_gb": 2.2, "params": "3.8B", "role": "chat", "note": "Strong for its size"},
{"name": "mistral:latest", "size_gb": 4.1, "params": "7B", "role": "chat/memory","note": "Good curator; runs on CPU/8GB+ RAM"},
{"name": "qwen2.5:7b", "size_gb": 4.7, "params": "7B", "role": "chat", "note": "Capable 7B"},
{"name": "llama3.1:8b", "size_gb": 4.7, "params": "8B", "role": "chat", "note": "Higher quality; CPU or 6GB+ GPU"},
{"name": "nomic-embed-text","size_gb": 0.27,"params": "—", "role": "embeddings", "note": "Required for recall / document RAG"},
{"name": "qwen2.5:14b", "size_gb": 9.0, "params": "14B", "role": "chat", "note": "High-end; needs a 12-16GB GPU or 24GB+ RAM"},
]