\n Professional Guidelines & Community\n

Empowering healthy smiles through evidence.

Access dentist-reviewed educational tracks, custom oral health routines, and a transparent, secure community forum run by dental professionals.

\n \n \n

\n Evidence-Based\n All guides referenced back to verified dental research.\n

\n Custom Care Routines\n Create structured brushing, flossing, or orthodontic plans.\n

\n Group Discussions\n Connect with our global community alongside simulated specialist panels.\n

\n\n \n

Essential Guides

\n \n

\n\n \n \n\n \n \n\n \n \n\n \n

\n \n

Standard scoring metrics represent compiled baseline performance parameters for user reference. Cleanliness maps to plaque reduction capability; protection maps to fluoride/HAP mineral coverage.

\n \n \n \n \n \n \n \n \n \n \n \n \n \n

Product	Type	Cleanliness Score	Protection	Value / Cost

\n\n \n \n\n \n \n\n \n \n\n \n\n`;\n\n return new Response(html, {\n headers: { \"Content-Type\": \"text/html; charset=utf-8\" }\n });\n}","children",[],[183],"5bd84edf-c102-4415-95b8-64268cde485d","frontend_styles.ts","01a020b8-e4a0-743e-8867-aae4d7aa810f",["D",1787255383090],"script","// frontend_styles.ts — CSS and shared HTML fragments for Open-Source-Lab\n\nexport const SHARED_CSS = `\n\n\n\n\n`;\n\nexport const CATALOG = [\n {n:'Llama-3.1-405B',src:'Meta',cat:'text',p:'405B',c:'128K',l:'Llama 3.1'},\n {n:'Llama-3.3-70B',src:'Meta',cat:'text',p:'70B',c:'128K',l:'Llama 3.3'},\n {n:'Qwen3-235B-A22B',src:'Alibaba',cat:'text',p:'235B/22B',c:'128K',l:'Apache 2.0'},\n {n:'Qwen3.5-27B',src:'Alibaba',cat:'text',p:'27B',c:'262K',l:'Apache 2.0'},\n {n:'Qwen2.5-72B',src:'Alibaba',cat:'text',p:'72B',c:'128K',l:'Qwen'},\n {n:'DeepSeek-V3.1',src:'DeepSeek',cat:'text',p:'671B/37B',c:'128K',l:'MIT'},\n {n:'DeepSeek-R1',src:'DeepSeek',cat:'text',p:'671B/37B',c:'128K',l:'MIT'},\n {n:'Mistral-Large-2',src:'Mistral AI',cat:'text',p:'123B',c:'128K',l:'MRL'},\n {n:'Mixtral-8x22B',src:'Mistral AI',cat:'text',p:'141B/39B',c:'64K',l:'Apache 2.0'},\n {n:'Mistral-NeMo',src:'Mistral/NVIDIA',cat:'text',p:'12B',c:'128K',l:'Apache 2.0'},\n {n:'Mistral-Small-3',src:'Mistral AI',cat:'text',p:'24B',c:'32K',l:'Apache 2.0'},\n {n:'Gemma-2-27B',src:'Google',cat:'text',p:'27B',c:'8K',l:'Gemma'},\n {n:'Gemma-2-9B',src:'Google',cat:'text',p:'9B',c:'8K',l:'Gemma'},\n {n:'Gemma-3-27B',src:'Google',cat:'text',p:'27B',c:'128K',l:'Gemma'},\n {n:'Phi-4',src:'Microsoft',cat:'text',p:'14B',c:'16K',l:'MIT'},\n {n:'Phi-3.5-MoE',src:'Microsoft',cat:'text',p:'42B/6.6B',c:'128K',l:'MIT'},\n {n:'Command-R+',src:'Cohere',cat:'text',p:'104B',c:'128K',l:'CC-BY-NC'},\n {n:'Command-R',src:'Cohere',cat:'text',p:'35B',c:'128K',l:'CC-BY-NC'},\n {n:'Hermes-3-405B',src:'Nous Research',cat:'text',p:'405B',c:'128K',l:'Llama 3.1'},\n {n:'Hermes-4-70B',src:'Nous Research',cat:'text',p:'70B',c:'128K',l:'Llama 3.1'},\n {n:'Yi-1.5-34B',src:'01.AI',cat:'text',p:'34B',c:'32K',l:'Apache 2.0'},\n {n:'InternLM2.5-20B',src:'Shanghai AI Lab',cat:'text',p:'20B',c:'256K',l:'Apache 2.0'},\n {n:'Falcon-2-11B',src:'TII',cat:'text',p:'11B',c:'8K',l:'Apache 2.0'},\n {n:'OLMo-2-13B',src:'Allen AI',cat:'text',p:'13B',c:'4K',l:'Apache 2.0'},\n {n:'DBRX',src:'Databricks',cat:'text',p:'132B/36B',c:'32K',l:'Databricks'},\n {n:'SmolLM2-1.7B',src:'HuggingFace',cat:'text',p:'1.7B',c:'8K',l:'Apache 2.0'},\n {n:'Trinity-Mini',src:'Open-Source-Lab',cat:'experimental',p:'Mini',c:'128K',l:'Open'},\n {n:'BitNet-b1.58-3B',src:'Microsoft',cat:'experimental',p:'3B',c:'4K',l:'MIT'},\n {n:'Mamba-2-2.7B',src:'Cartesia/CMU',cat:'experimental',p:'2.7B',c:'1M',l:'Apache 2.0'},\n {n:'RWKV-6-World-7B',src:'RWKV Foundation',cat:'experimental',p:'7B',c:'Inf',l:'Apache 2.0'},\n {n:'Samba-CoE-v0.2',src:'SambaNova',cat:'experimental',p:'40B/8B',c:'256K',l:'Apache 2.0'},\n {n:'Chameleon-7B',src:'Meta',cat:'experimental',p:'7B',c:'8K',l:'Llama 3.1'},\n {n:'Laguna-M.1',src:'Poolside',cat:'code',p:'—',c:'128K',l:'Hosted'},\n {n:'GLM-5.1',src:'Zhipu AI',cat:'code',p:'754B',c:'205K',l:'MIT'},\n {n:'Minimax-M2.7',src:'MiniMax',cat:'code',p:'—',c:'205K',l:'Open'},\n {n:'MiMo-V2-Flash',src:'Xiaomi MiMo',cat:'code',p:'309B/15B',c:'256K',l:'Apache 2.0'},\n {n:'Qwen2.5-Coder-32B',src:'Alibaba',cat:'code',p:'32B',c:'128K',l:'Apache 2.0'},\n {n:'DeepSeek-Coder-V2',src:'DeepSeek',cat:'code',p:'236B/21B',c:'128K',l:'MIT'},\n {n:'StarCoder2-15B',src:'BigCode',cat:'code',p:'15B',c:'16K',l:'OpenRAIL'},\n {n:'Mistral-Small-4',src:'Mistral AI',cat:'multimodal',p:'24B',c:'128K',l:'Apache 2.0'},\n {n:'Nova-Lite-2',src:'Amazon',cat:'multimodal',p:'—',c:'1M',l:'Hosted'},\n {n:'Qwen2.5-VL-72B',src:'Alibaba',cat:'vision',p:'72B',c:'128K',l:'Apache 2.0'},\n {n:'Pixtral-12B',src:'Mistral AI',cat:'vision',p:'12B',c:'128K',l:'Apache 2.0'},\n {n:'Llama-3.2-90B-Vision',src:'Meta',cat:'vision',p:'90B',c:'128K',l:'Llama 3.2'},\n {n:'InternVL2.5-78B',src:'OpenGVLab',cat:'vision',p:'78B',c:'32K',l:'MIT'},\n {n:'Whisper-Large-v3',src:'OpenAI',cat:'audio',p:'1.55B',c:'—',l:'MIT'},\n {n:'Qwen2-Audio',src:'Alibaba',cat:'audio',p:'7B',c:'—',l:'Apache 2.0'},\n {n:'BGE-M3',src:'BAAI',cat:'embed',p:'560M',c:'8192',l:'MIT'},\n {n:'Nomic-Embed-v1.5',src:'Nomic AI',cat:'embed',p:'137M',c:'8192',l:'Apache 2.0'},\n {n:'F5-TTS',src:'F5-TTS Team',cat:'audio',p:'335M',c:'—',l:'MIT'},\n {n:'Moshi-v1.0',src:'Kyutai',cat:'audio',p:'7B',c:'—',l:'CC-BY 4.0'},\n];\n\nexport const CAT_NAMES: Record = {\n all:'All',text:'Language',code:'Code',vision:'Vision',\n audio:'Audio',multimodal:'Multimodal',embed:'Embeddings',experimental:'Experimental'\n};\n\nexport const BENCH_DATA = [\n {n:'Qwen-3-235B-A22B',o:'Alibaba/Qwen',cat:'text',os:true,mmlu:87,gpqa:71,he:88,math:86},\n {n:'Hermes-3-405B',o:'Nous Research',cat:'text',os:true,mmlu:86,gpqa:59,he:86,math:73},\n {n:'Hermes-4-70B',o:'Nous Research',cat:'text',os:true,mmlu:82,gpqa:55,he:82,math:70},\n {n:'Phi-4',o:'Microsoft Research',cat:'text',os:true,mmlu:84.8,gpqa:56.1,he:82.6,math:80.4},\n {n:'Qwen3.5-27B',o:'Alibaba/Qwen',cat:'text',os:true,mmlu:83,gpqa:60,he:89,math:82},\n {n:'DeepSeek-V3.1',o:'DeepSeek',cat:'text',os:true,mmlu:88.5,gpqa:59.1,he:90.2,math:90.2},\n {n:'Mistral-NeMo',o:'Mistral/NVIDIA',cat:'text',os:true,mmlu:68,gpqa:40,he:70,math:55},\n {n:'GLM-5.1',o:'Zhipu AI',cat:'code',os:true,mmlu:88,gpqa:72,he:94,math:88},\n {n:'MiMo-V2-Flash',o:'Xiaomi MiMo',cat:'code',os:true,mmlu:82,gpqa:63,he:89,math:84},\n {n:'Qwen2.5-Coder-32B',o:'Alibaba/Qwen',cat:'code',os:true,mmlu:75,gpqa:48,he:92.7,math:75},\n {n:'DeepSeek-Coder-V2',o:'DeepSeek',cat:'code',os:true,mmlu:79.2,gpqa:50,he:90.2,math:75.7},\n {n:'Qwen2.5-VL-72B',o:'Alibaba/Qwen',cat:'multimodal',os:true,mmlu:79,gpqa:51,he:80,math:68},\n {n:'Mistral-Small-4',o:'Mistral AI',cat:'multimodal',os:true,mmlu:78,gpqa:52,he:84,math:70},\n {n:'Nova-Lite-2',o:'Amazon',cat:'multimodal',os:false,mmlu:80,gpqa:50,he:82,math:72},\n {n:'Llama-3.1-405B',o:'Meta',cat:'ref',os:true,mmlu:88.6,gpqa:51.1,he:89,math:73.8},\n {n:'Qwen2.5-72B',o:'Alibaba/Qwen',cat:'ref',os:true,mmlu:86.1,gpqa:49,he:86.6,math:83.1},\n {n:'Gemma-2-27B',o:'Google',cat:'ref',os:true,mmlu:75.2,gpqa:42,he:72,math:55},\n {n:'Mixtral-8x22B',o:'Mistral AI',cat:'ref',os:true,mmlu:77.8,gpqa:44,he:76,math:54},\n].map(r => ({...r, avg: Math.round(((r.mmlu+r.gpqa+r.he+r.math)/4)*10)/10}));",[],[191],"3281360f-b4a6-4cce-81cc-c897731b5cdf","frontend_html.ts","01a020b8-e4a0-743e-8867-a60c3ff7073c",["D",1787255383090],"// frontend_html.ts — HTML page builder for Open-Source-Lab on val.town\nimport { SHARED_CSS, CATALOG, CAT_NAMES, BENCH_DATA } from \"./frontend_styles.ts\";\n\nfunction esc(s: string) {\n return s.replace(/&/g,'&').replace(//g,'>').replace(/\"/g,'"');\n}\n\nfunction badge(tag: string) {\n const map: Record = {\n open:'badge-open',new:'badge-new',exp:'badge-exp',hot:'badge-hot',local:'badge-local'\n };\n const labels: Record = {\n open:'Open',new:'New',exp:'Experimental',hot:'Popular',local:'Local'\n };\n if (!map[tag]) return '';\n return `${labels[tag]}`;\n}\n\nfunction modelCard(m: any) {\n const init = m.n.replace(/[^A-Za-z0-9]/g,'').slice(0,2).toUpperCase();\n return `

${init}

${['open','new','hot','local'].filter(() => Math.random()>.6).map(badge).join('')}

${esc(m.n)}

${esc(m.src)}

${esc(m.p)}Params

${esc(m.c)}Context

${esc(m.l)}License

`;\n}\n\nfunction sectionCards(cat: string, title: string, limit = 6) {\n const items = CATALOG.filter(m => m.cat === cat).slice(0, limit);\n return `

${title}

\n View all\n

${items.map(modelCard).join('')}

`;\n}\n\nfunction libCards() {\n return CATALOG.map(m => {\n const init = m.n.replace(/[^A-Za-z0-9]/g,'').slice(0,2).toUpperCase();\n return `

${esc(m.n)}

${esc(m.src)}

\n ${CAT_NAMES[m.cat]||m.cat}\n

${esc(m.p)}Params

${esc(m.c)}Context

${esc(m.l)}License

`;\n }).join('');\n}\n\nfunction benchRows() {\n return BENCH_DATA.map(r => {\n const typePill = r.cat==='ref'\n ? `Baseline`\n : (r.os ? `Open` : `Hosted`);\n const cell = (v: number) => `${v.toFixed(1)}

`;\n return `\n

${esc(r.n)}

${esc(r.o)}

\n ${typePill}\n ${cell(r.mmlu)}${cell(r.gpqa)}${cell(r.he)}${cell(r.math)}${cell(r.avg)}\n`;\n }).join('');\n}\n\nexport function buildPage(chatModels: Array<{id:string,name:string,org:string}>) {\n const modelOptions = chatModels.map(m =>\n ``\n ).join('');\n\n const poeLinks = [\n {n:'Qwen3.5-27B',d:'Alibaba\\'s dense hybrid reasoning model.',url:'https://poe.com/Qwen3.5-27B'},\n {n:'Hermes-3-405B',d:'Nous Research instruction-following flagship.',url:'https://poe.com/Hermes-3-405B'},\n {n:'DeepSeek-V3',d:'671B MoE with advanced mathematics reasoning.',url:'https://poe.com/DeepSeek-V3'},\n {n:'Phi-4',d:'Microsoft small-footprint reasoning model.',url:'https://poe.com/Phi-4-OSL'},\n {n:'Mistral-NeMo',d:'12B collaborative build, multilingual tasks.',url:'https://poe.com/Mistral-NeMo'},\n {n:'Qwen2.5-Coder-32B',d:'Baseline workhorse for open coding.',url:'https://poe.com/Qwen2.5-Coder-32B'},\n ];\n\n return `\n\n\n\n\nOpen-Source-Lab — Local Model Catalogue\n${SHARED_CSS}\n\n\n\n\n\n\n\n

Model Catalogue

A place
to test
local models

Curated local and open-source AI models. From community workhorses to experimental frontier architectures.

\n Browse 50 Models\n Try Live Chat\n

Test Local ModelsTry models instantly without installing them.

Check the SpecsCompare parameters, context sizes, find your niche.

Exchange KnowledgeDiscuss hardware with open source enthusiasts.

\n\n\n

Model of the Month

Spotlight — June 2026

GLM-5.1

\n Open\n New\n Popular\n

Zhipu AI's agentic coding flagship posts SOTA on SWE-bench Pro and sustains thousands of tool calls. MIT-licensed, fully self-hostable for end-to-end project execution, debugging, and log analysis at scale.

754BParams

205KContext

MITLicense

MoEArch

\n Test Model\n Try Live Chat\n

\n\n\n

\n \n \n \n \n \n \n \n \n

\n\n\n

\n Key:\n

OpenApache / MIT / similar

NewReleased last 90 days

ExperimentalEarly / research stage

PopularCommunity favourite

LocalConsumer hardware

\n\n${sectionCards('text','Language & Reasoning Models')}\n${sectionCards('code','Code & Engineering Models')}\n${sectionCards('vision','Vision & Multimodal',7)}\n${sectionCards('audio','Audio & Embedding Models')}\n${sectionCards('experimental','Experimental & Research')}\n\n\n

Ready to Go Local?

Start running open models today. No API keys, no subscriptions, no data leaving your machine.

\n Try Live Chat\n View Benchmarks\n

\n\n\n\n\n\n\n\n\n\n\n\n\n\n

\n \n

\n \n \n \n

1. Pick an inference engine Start here

Your engine determines speed, hardware support and ease of use. The most popular options in 2026:

Ollama — easiest entry point; one-line model pulls and a built-in API. Great for beginners and Mac users.
llama.cpp — C/C++ core powering most local tooling. Runs GGUF models on CPU, CUDA, Metal, Vulkan and ROCm.
vLLM — high-throughput GPU serving with PagedAttention; ideal for multi-user or production workloads.
LM Studio — polished desktop GUI for browsing, downloading and chatting with GGUF models, no terminal needed.

$ ollama run qwen3.5:27b\n$ ./llama-server -m model.gguf -ngl 99 -c 8192

2. Understand quantization

Quantization shrinks weights from 16-bit down to fewer bits, trading a little quality for big memory savings.

Q8_0 — near-lossless, ~1 byte/param. Use when you have the VRAM.
Q4_K_M — sweet spot. ~4.5 bits/param with minimal quality loss; the default most people run.
Q3 / Q2 — emergency tier for fitting big models on small cards; noticeable degradation.
Prefer a larger model at Q4 over a smaller at Q8 when quality matters.

3. Memory math (VRAM estimate)

At Q4 you need roughly 0.55 GB per billion parameters, plus context overhead.

7B @ Q4 → ~5 GB · fits an 8 GB card.
14B @ Q4 → ~9 GB · fits 12 GB.
32B @ Q4 → ~20 GB · needs 24 GB (RTX 4090 / 3090).
70B @ Q4 → ~42 GB · dual-GPU or 48 GB+, or offload to RAM.

Long context inflates the KV cache; a 128K window can add several GB on top.

4. Hardware tips

Apple Silicon shines for local LLMs thanks to unified memory — M-series with 64GB+ runs large MoE models comfortably.
NVIDIA remains fastest via CUDA; the 3090/4090's 24GB is the consumer sweet spot.
CPU + RAM offload works for big models but is slow; expect single-digit tokens/sec.
MoE models activate only a fraction of params per token, so they run far faster than their total size suggests.

5. Performance tips

Set -ngl / GPU layers as high as your VRAM allows to offload the whole model to GPU.
Lower temperature (0.1–0.3) for code and factual tasks; raise it (0.7–1.0) for creative writing.
Enable flash attention and KV-cache quantization to fit longer contexts.
For reasoning models, allow plenty of output tokens so chain-of-thought isn't cut off.

\n\n\n

\n \n

\n \n \n \n

\n \n \n \n \n \n

\n \n \n \n \n \n \n \n \n \n \n ${benchRows()}\n

Model	Type	MMLU	GPQA	HumanEval	MATH	Avg

MMLU = general knowledge · GPQA = graduate-level science (Diamond) · HumanEval = code generation · MATH = competition mathematics. Scores are representative composite figures from public reports — verify against primary sources before citing.

\n\n\n\n\n\n\n\n\n\n\n\n\n\n`;\n}",[],[199],"6ba539e0-6bb2-4048-b054-6f2516addae9","db.ts","01a020b8-e4a0-743e-8867-a369617d541d",["D",1787255383090],"// db.ts — SQLite schema and helpers for Open-Source-Lab on val.town\n// Import from val.town's sqlite module in your actual val:\n// import { sqlite } from \"https://esm.town/v/std/sqlite\";\n\nexport async function initDB(sqlite: any) {\n await sqlite.execute(`\n CREATE TABLE IF NOT EXISTS users (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n username TEXT UNIQUE NOT NULL COLLATE NOCASE,\n password_hash TEXT NOT NULL,\n created_at INTEGER DEFAULT (unixepoch()),\n is_admin INTEGER DEFAULT 0\n )\n `);\n await sqlite.execute(`\n CREATE TABLE IF NOT EXISTS sessions (\n token TEXT PRIMARY KEY,\n user_id INTEGER NOT NULL,\n created_at INTEGER DEFAULT (unixepoch()),\n expires_at INTEGER NOT NULL\n )\n `);\n await sqlite.execute(`\n CREATE TABLE IF NOT EXISTS chat_usage (\n id INTEGER PRIMARY KEY AUTOINCREMENT,\n user_id INTEGER NOT NULL,\n model TEXT NOT NULL,\n input_chars INTEGER DEFAULT 0,\n output_chars INTEGER DEFAULT 0,\n created_at INTEGER DEFAULT (unixepoch())\n )\n `);\n // Ensure roboohood is admin\n await sqlite.execute(`\n INSERT OR IGNORE INTO users (username, password_hash, is_admin)\n VALUES ('roboohood', '${await hashPassword(\"admin123\")}', 1)\n `);\n}\n\nexport async function hashPassword(password: string): Promise {\n const encoder = new TextEncoder();\n const data = encoder.encode(password + \"osl_salt_2026\");\n const hashBuffer = await crypto.subtle.digest(\"SHA-256\", data);\n const hashArray = Array.from(new Uint8Array(hashBuffer));\n return hashArray.map(b => b.toString(16).padStart(2, \"0\")).join(\"\");\n}\n\nexport async function createUser(sqlite: any, username: string, password: string) {\n if (!/^[a-zA-Z0-9_]{3,32}$/.test(username)) {\n throw new Error(\"Username must be 3–32 characters: letters, numbers, underscores only.\");\n }\n if (password.length < 6) {\n throw new Error(\"Password must be at least 6 characters.\");\n }\n const hash = await hashPassword(password);\n const isAdmin = username.toLowerCase() === \"roboohood\" ? 1 : 0;\n try {\n await sqlite.execute(\n `INSERT INTO users (username, password_hash, is_admin) VALUES (?, ?, ?)`,\n [username, hash, isAdmin]\n );\n } catch (e: any) {\n if (e.message?.includes(\"UNIQUE\")) throw new Error(\"Username already taken.\");\n throw e;\n }\n}\n\nexport async function verifyUser(sqlite: any, username: string, password: string) {\n const hash = await hashPassword(password);\n const rows = await sqlite.execute(\n `SELECT id, username, is_admin FROM users WHERE username = ? COLLATE NOCASE AND password_hash = ?`,\n [username, hash]\n );\n if (!rows.rows || rows.rows.length === 0) return null;\n const [id, uname, isAdmin] = rows.rows[0];\n return { id, username: uname, isAdmin: !!isAdmin };\n}\n\nexport async function createSession(sqlite: any, userId: number): Promise {\n const token = crypto.randomUUID() + crypto.randomUUID();\n const expiresAt = Math.floor(Date.now() / 1000) + 60 * 60 * 24 * 7; // 7 days\n await sqlite.execute(\n `INSERT INTO sessions (token, user_id, expires_at) VALUES (?, ?, ?)`,\n [token, userId, expiresAt]\n );\n return token;\n}\n\nexport async function getSession(sqlite: any, token: string) {\n if (!token) return null;\n const now = Math.floor(Date.now() / 1000);\n const rows = await sqlite.execute(\n `SELECT s.user_id, u.username, u.is_admin\n FROM sessions s JOIN users u ON u.id = s.user_id\n WHERE s.token = ? AND s.expires_at > ?`,\n [token, now]\n );\n if (!rows.rows || rows.rows.length === 0) return null;\n const [userId, username, isAdmin] = rows.rows[0];\n return { userId, username, isAdmin: !!isAdmin };\n}\n\nexport async function deleteSession(sqlite: any, token: string) {\n await sqlite.execute(`DELETE FROM sessions WHERE token = ?`, [token]);\n}\n\nexport async function changePassword(sqlite: any, userId: number, oldPass: string, newPass: string) {\n // Verify old password\n const oldHash = await hashPassword(oldPass);\n const check = await sqlite.execute(\n `SELECT id FROM users WHERE id = ? AND password_hash = ?`,\n [userId, oldHash]\n );\n if (!check.rows || check.rows.length === 0) throw new Error(\"Current password is incorrect.\");\n if (newPass.length < 6) throw new Error(\"New password must be at least 6 characters.\");\n const newHash = await hashPassword(newPass);\n await sqlite.execute(`UPDATE users SET password_hash = ? WHERE id = ?`, [newHash, userId]);\n}\n\nexport async function logChatUsage(\n sqlite: any,\n userId: number,\n model: string,\n inputChars: number,\n outputChars: number\n) {\n await sqlite.execute(\n `INSERT INTO chat_usage (user_id, model, input_chars, output_chars) VALUES (?, ?, ?, ?)`,\n [userId, model, inputChars, outputChars]\n );\n}\n\nexport async function getAllUsers(sqlite: any) {\n const rows = await sqlite.execute(\n `SELECT u.id, u.username, u.is_admin, u.created_at,\n COALESCE(SUM(c.input_chars),0) AS total_in,\n COALESCE(SUM(c.output_chars),0) AS total_out\n FROM users u LEFT JOIN chat_usage c ON c.user_id = u.id\n GROUP BY u.id ORDER BY u.created_at DESC`\n );\n if (!rows.rows) return [];\n return rows.rows.map(([id, username, isAdmin, createdAt, totalIn, totalOut]: any) => ({\n id, username, isAdmin: !!isAdmin, createdAt, totalIn, totalOut\n }));\n}\n\nexport async function deleteUser(sqlite: any, userId: number) {\n await sqlite.execute(`DELETE FROM sessions WHERE user_id = ?`, [userId]);\n await sqlite.execute(`DELETE FROM chat_usage WHERE user_id = ?`, [userId]);\n await sqlite.execute(`DELETE FROM users WHERE id = ?`, [userId]);\n}\n\nexport async function getUserStats(sqlite: any, userId: number) {\n const rows = await sqlite.execute(\n `SELECT COALESCE(SUM(input_chars),0), COALESCE(SUM(output_chars),0)\n FROM chat_usage WHERE user_id = ?`,\n [userId]\n );\n if (!rows.rows || rows.rows.length === 0) return { totalIn: 0, totalOut: 0 };\n const [totalIn, totalOut] = rows.rows[0];\n return { totalIn, totalOut };\n}",[],[206],"1abc28a0-c941-42d3-bf71-7ce1ab7c7ad1","api.ts","01a020b8-e4a0-743e-8867-9e3a547dc533",["D",1787255383090],"// api.ts — Route handlers for Open-Source-Lab on val.town\n// All functions receive (req, sqlite) and return Response\n\nimport {\n createUser, verifyUser, createSession, getSession,\n deleteSession, changePassword, logChatUsage,\n getAllUsers, deleteUser, getUserStats\n} from \"./db.ts\";\n\n// ─── helpers ──────────────────────────────────────────────────────────────────\n\nfunction json(data: any, status = 200) {\n return new Response(JSON.stringify(data), {\n status,\n headers: { \"Content-Type\": \"application/json\" }\n });\n}\n\nfunction err(msg: string, status = 400) {\n return json({ error: msg }, status);\n}\n\nfunction getToken(req: Request): string {\n const cookie = req.headers.get(\"cookie\") || \"\";\n const match = cookie.match(/osl_session=([^;]+)/);\n return match ? match[1] : \"\";\n}\n\n// ─── Chat models ──────────────────────────────────────────────────────────────\n\nexport const CHAT_MODELS = [\n // Groq — free tier, fast\n { id: \"llama-3.1-8b-instant\", name: \"Llama 3.1 8B (Groq)\", provider: \"groq\", org: \"Meta\" },\n { id: \"llama-3.3-70b-versatile\", name: \"Llama 3.3 70B (Groq)\", provider: \"groq\", org: \"Meta\" },\n { id: \"gemma2-9b-it\", name: \"Gemma 2 9B (Groq)\", provider: \"groq\", org: \"Google\" },\n { id: \"mistral-saba-24b\", name: \"Mistral Saba 24B (Groq)\", provider: \"groq\", org: \"Mistral AI\" },\n { id: \"qwen-qwq-32b\", name: \"QwQ 32B (Groq)\", provider: \"groq\", org: \"Alibaba\" },\n { id: \"deepseek-r1-distill-llama-70b\", name: \"DeepSeek R1 Distill 70B (Groq)\", provider: \"groq\", org: \"DeepSeek\" },\n // Mistral — open weights on their platform\n { id: \"open-mistral-nemo\", name: \"Mistral NeMo (Mistral)\", provider: \"mistral\", org: \"Mistral AI\" },\n { id: \"open-mistral-7b\", name: \"Mistral 7B (Mistral)\", provider: \"mistral\", org: \"Mistral AI\" },\n { id: \"open-mixtral-8x7b\", name: \"Mixtral 8x7B (Mistral)\", provider: \"mistral\", org: \"Mistral AI\" },\n { id: \"mistral-small-latest\", name: \"Mistral Small (Mistral)\", provider: \"mistral\", org: \"Mistral AI\" },\n];\n\n// ─── POST /api/register ───────────────────────────────────────────────────────\n\nexport async function handleRegister(req: Request, sqlite: any): Promise {\n let body: any;\n try { body = await req.json(); } catch { return err(\"Invalid JSON\"); }\n const { username, password } = body;\n if (!username || !password) return err(\"Username and password required.\");\n try {\n await createUser(sqlite, username, password);\n const user = await verifyUser(sqlite, username, password);\n if (!user) return err(\"Registration failed.\");\n const token = await createSession(sqlite, user.id);\n const resp = json({ ok: true, username: user.username, isAdmin: user.isAdmin });\n resp.headers.set(\"Set-Cookie\", `osl_session=${token}; Path=/; HttpOnly; SameSite=Lax; Max-Age=604800`);\n return resp;\n } catch (e: any) {\n return err(e.message);\n }\n}\n\n// ─── POST /api/login ─────────────────────────────────────────────────────────\n\nexport async function handleLogin(req: Request, sqlite: any): Promise {\n let body: any;\n try { body = await req.json(); } catch { return err(\"Invalid JSON\"); }\n const { username, password } = body;\n if (!username || !password) return err(\"Username and password required.\");\n const user = await verifyUser(sqlite, username, password);\n if (!user) return err(\"Invalid username or password.\", 401);\n const token = await createSession(sqlite, user.id);\n const resp = json({ ok: true, username: user.username, isAdmin: user.isAdmin });\n resp.headers.set(\"Set-Cookie\", `osl_session=${token}; Path=/; HttpOnly; SameSite=Lax; Max-Age=604800`);\n return resp;\n}\n\n// ─── POST /api/logout ────────────────────────────────────────────────────────\n\nexport async function handleLogout(req: Request, sqlite: any): Promise {\n const token = getToken(req);\n if (token) await deleteSession(sqlite, token);\n const resp = json({ ok: true });\n resp.headers.set(\"Set-Cookie\", \"osl_session=; Path=/; Max-Age=0\");\n return resp;\n}\n\n// ─── GET /api/me ─────────────────────────────────────────────────────────────\n\nexport async function handleMe(req: Request, sqlite: any): Promise {\n const token = getToken(req);\n const session = await getSession(sqlite, token);\n if (!session) return json({ user: null });\n const stats = await getUserStats(sqlite, session.userId);\n return json({ user: { username: session.username, isAdmin: session.isAdmin, ...stats } });\n}\n\n// ─── POST /api/change-password ───────────────────────────────────────────────\n\nexport async function handleChangePassword(req: Request, sqlite: any): Promise {\n const token = getToken(req);\n const session = await getSession(sqlite, token);\n if (!session) return err(\"Not logged in.\", 401);\n let body: any;\n try { body = await req.json(); } catch { return err(\"Invalid JSON\"); }\n const { oldPassword, newPassword } = body;\n if (!oldPassword || !newPassword) return err(\"Both passwords required.\");\n try {\n await changePassword(sqlite, session.userId, oldPassword, newPassword);\n return json({ ok: true });\n } catch (e: any) {\n return err(e.message);\n }\n}\n\n// ─── POST /api/chat ──────────────────────────────────────────────────────────\n\nexport async function handleChat(req: Request, sqlite: any, env: any): Promise {\n const token = getToken(req);\n const session = await getSession(sqlite, token);\n if (!session) return err(\"You must be logged in to use the chat.\", 401);\n\n let body: any;\n try { body = await req.json(); } catch { return err(\"Invalid JSON\"); }\n\n const { modelId, messages } = body;\n if (!modelId || !Array.isArray(messages)) return err(\"modelId and messages required.\");\n\n const model = CHAT_MODELS.find(m => m.id === modelId);\n if (!model) return err(\"Unknown model.\");\n\n // Count input chars\n const inputChars = messages.reduce((acc: number, m: any) =>\n acc + String(m.content || \"\").length, 0);\n\n try {\n let response: Response;\n\n if (model.provider === \"groq\") {\n response = await fetch(\"https://api.groq.com/openai/v1/chat/completions\", {\n method: \"POST\",\n headers: {\n \"Content-Type\": \"application/json\",\n \"Authorization\": `Bearer ${env.GROQ_API_KEY}`\n },\n body: JSON.stringify({ model: modelId, messages, max_tokens: 1024 })\n });\n } else if (model.provider === \"mistral\") {\n response = await fetch(\"https://api.mistral.ai/v1/chat/completions\", {\n method: \"POST\",\n headers: {\n \"Content-Type\": \"application/json\",\n \"Authorization\": `Bearer ${env.MISTRAL_API_KEY}`\n },\n body: JSON.stringify({ model: modelId, messages, max_tokens: 1024 })\n });\n } else {\n return err(\"Unknown provider.\");\n }\n\n if (!response.ok) {\n const errText = await response.text();\n return err(`Model API error: ${response.status} ${errText}`, 502);\n }\n\n const data = await response.json();\n const reply = data.choices?.[0]?.message?.content || \"(No response)\";\n const outputChars = reply.length;\n\n await logChatUsage(sqlite, session.userId, modelId, inputChars, outputChars);\n\n return json({ reply, model: model.name });\n } catch (e: any) {\n return err(`Request failed: ${e.message}`, 502);\n }\n}\n\n// ─── GET /api/admin/users ─────────────────────────────────────────────────────\n\nexport async function handleAdminUsers(req: Request, sqlite: any): Promise {\n const token = getToken(req);\n const session = await getSession(sqlite, token);\n if (!session?.isAdmin) return err(\"Admin only.\", 403);\n const users = await getAllUsers(sqlite);\n return json({ users });\n}\n\n// ─── DELETE /api/admin/users/:id ─────────────────────────────────────────────\n\nexport async function handleAdminDeleteUser(req: Request, sqlite: any, userId: number): Promise {\n const token = getToken(req);\n const session = await getSession(sqlite, token);\n if (!session?.isAdmin) return err(\"Admin only.\", 403);\n if (session.userId === userId) return err(\"Cannot delete yourself.\");\n await deleteUser(sqlite, userId);\n return json({ ok: true });\n}\n\n// ─── GET /api/models ─────────────────────────────────────────────────────────\n\nexport async function handleModels(_req: Request): Promise {\n return json({ models: CHAT_MODELS });\n}",[],[213],["D",1787255383090],[],[18],["D",1787255383090],["D",1789520505489],{"_11":12,"_126":70},{"_11":12,"_121":122,"_69":70,"_30":68,"_123":124,"_125":24},{"_38":39,"_13":16,"_32":35,"_111":72,"_112":62,"_113":39,"_114":76,"_115":116,"_117":118,"_119":14,"_120":12},"ctx",{"_51":-5,"_147":260,"_261":262,"_263":264,"_265":24,"_266":24,"_267":24,"_268":85,"_269":-5,"_270":-7,"_271":-7,"_272":26,"_273":24,"_274":24,"_275":-5,"_276":24,"_277":24,"_278":24,"_279":-5},"ENV",{"_231":232,"_233":234,"_235":16,"_236":237,"_238":239,"_240":241,"_242":243,"_244":245,"_246":247,"_248":249,"_250":251,"_252":253,"_254":255,"_256":257,"_258":259},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","fdb67791413f4386b4dbf18f354029a617f6160f","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]

Empowering healthy smiles through evidence.

Essential Guides

All Educational Tracks

Your Custom Care Plans

Recoverable Trash

${title}

A place
to test
local models

Model of the Month

Ready to Go Local?

Empowering healthy smiles through evidence.

Essential Guides

All Educational Tracks

Your Custom Care Plans

Recoverable Trash

${title}

A placeto testlocal models

Model of the Month

Ready to Go Local?

1. Pick an inference engine Start here

2. Understand quantization

3. Memory math (VRAM estimate)

4. Hardware tips

5. Performance tips

A place
to test
local models

History · Readme.txt