SOURCE_HUB

\n {[...MESH_REGISTRY.entries()].map(([id, data]) => (\n

● {id.toUpperCase()}

{JSON.stringify(data, null, 2)}

\n ))}\n

\n\n

\n \n \n \n ,\n );\n});\n\n// 4. HARDENED DISPATCH\napp.post(\"/dispatch\", async (c) => {\n const { cmd } = await c.req.parseBody();\n\n // Brutality: Retry 3 times with exponential backoff before admitting defeat\n const result = await pRetry(async () => {\n // Selection logic based on MESH_REGISTRY health\n return \"RESULT_STUB\";\n }, { retries: 3 });\n\n return c.text(result);\n});\n\nexport default app.fetch;","deleted","label","children",[],[158],{"_105":179,"_96":180,"_159":181,"_161":-5,"_162":144,"_163":182,"_165":166,"_107":183,"_101":168,"_169":170,"_171":184,"_173":22,"_174":-5,"_175":185,"_99":186},"69a6800c-f689-11f0-9211-42dde27851f2","Brutal2.ts","28e00b42-f689-11f0-af96-42dde27851f2",134,["D",1768973049189],"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\n\nconst app = new Hono();\n\n// PRE-FILLED QUANTUM CHAOS PROMPT\nconst QUANTUM_PROMPT =\n `Analyze the non-local wave-function collapse in a double-slit interference pattern modulated by a stochastic gravitational background. Calculate the decoherence rate \\Gamma for a particle of mass M passing through a superposition of space-time metrics g_{\\mu\\nu} and \\tilde{g}_{\\mu\\nu}.`;\n\nconst NODES = [\n {\n id: \"POLLINATIONS_TEXT\",\n url: \"https://text.pollinations.ai/\",\n type: \"raw\",\n },\n {\n id: \"POLLINATIONS_V1\",\n url: \"https://gen.pollinations.ai/v1/chat/completions\",\n type: \"openai\",\n },\n { id: \"PUTER_V2\", url: \"https://api.puter.com/v2/chat\", type: \"openai\" },\n {\n id: \"AVALAI_LATEST\",\n url: \"https://api.avalai.ir/v1/chat/completions\",\n type: \"openai\",\n },\n {\n id: \"HF_INFERENCE\",\n url:\n \"https://api-inference.huggingface.co/models/mistralai/Mistral-7B-v0.3\",\n type: \"hf\",\n },\n];\n\napp.get(\"/\", (c) => {\n return c.html(\n \n \n VAL_TOWN_STRESS_TEST\n \n \n \n

\n SHADOW_STRESS_V1\n

\n // ENVIRONMENT: VAL_TOWN_DENO // PROTOCOL: KEYLESS_PULSE\n

\n\n {NODES.map((node) => (\n

\n NODE_ID: {node.id}\n

\n \n \n EXECUTE_PROBE\n \n

READY_FOR_INJECTION

\n ))}\n\n 200_OK' : '' + res.status + '_FAIL';\n const rawText = await res.text();\n\n mon.innerHTML = \\`\n

STATUS: \\${statusBadge} | LATENCY: \\${latency}ms

\\${rawText.substring(0, 600)}...

\n \\`;\n } catch (e) {\n mon.innerHTML = 'TERMINAL_ERR ' + e.message;\n }\n }\n `,\n }}\n />\n \n ,\n );\n});\n\nexport default app.fetch;",[],[180],{"_105":188,"_96":189,"_159":190,"_161":-5,"_162":144,"_163":191,"_165":166,"_107":192,"_101":193,"_169":170,"_171":194,"_173":22,"_174":-5,"_175":195,"_99":196},"c730143c-f68d-11f0-8eec-42dde27851f2","Brutalist4.ts","bec47054-f68d-11f0-99a5-42dde27851f2",150,["D",1768974924103],3,"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\n\nconst app = new Hono();\n\napp.get(\"/\", (c) => {\n return c.html(\n \n \n POLLINATIONS_ULTIMATE_V10\n \n \n \n

POLLINATIONS_ULTIMATE_V10

\n \n

\n \n \n

\n\n

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

\\${children}

\n \n \n \n `}\";\n };\n\n app.get(\"/\", async (c) => {\n const user = await scope.Auth.getSession(c);\n const flowRes = await sqlite.execute(\"SELECT name, idea FROM \" + TABLES.FLOW + \" ORDER BY ts DESC LIMIT 5\");\n \n const fragments = flowRes.rows.map(f => \"${`\n

\\${f[0]}

\\${f[1]}

\n `}\").join('');\n\n return c.html(await scope.UI.Layout({\n title: \"Nexus\", user,\n children: \"${`\n

\n \n

\\${html.raw(fragments)}

\n \n

\n `}\"\n }));\n });\n\n app.get(\"/s/:slug/:folder/:page\", async (c) => {\n const user = await scope.Auth.getSession(c);\n const path = c.req.path;\n const nodeRes = await sqlite.execute({ sql: \"SELECT html, title FROM \" + TABLES.NODES + \" WHERE path = ? AND generated = 1\", args: [path] });\n if (!nodeRes.rows[0]) return c.redirect(\"/\");\n \n return c.html(await scope.UI.Layout({\n title: nodeRes.rows[0][1], user,\n children: \"${`\n

NODE // \\${nodeRes.rows[0][1]}

\\${html.raw(nodeRes.rows[0][0])}

\n `}\"\n }));\n });\n `;\n\n await sqlite.execute({\n sql: `INSERT OR REPLACE INTO ${TABLES.KERNEL} (name, code) VALUES (?, ?)`,\n args: [\"main_logic\", logic],\n });\n },\n\n formatSlug: (s) =>\n (s || \"void\").toLowerCase().trim().replace(/[^\\w\\s-]/g, \"\").replace(\n /[\\s_-]+/g,\n \"-\",\n ).replace(/^-+|-+$/g, \"\"),\n};\n\nconst AI = {\n genAI: new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\"),\n async query(system, prompt) {\n try {\n const model = this.genAI.getGenerativeModel({\n model: \"gemini-1.5-flash\",\n systemInstruction: system,\n });\n const result = await model.generateContent({\n contents: [{ role: \"user\", parts: [{ text: prompt }] }],\n generationConfig: {\n responseMimeType: system.includes(\"JSON\")\n ? \"application/json\"\n : \"text/plain\",\n },\n });\n return (await result.response).text().replace(\n /```(json|html|javascript)?/g,\n \"\",\n ).replace(/```/g, \"\").trim();\n } catch (e) {\n return \"ERROR\";\n }\n },\n};\n\nexport default async function (req: Request) {\n // await DB.init();\n const runnerApp = new Hono();\n const AsyncFunction = Object.getPrototypeOf(async function () {}).constructor;\n\n const scope = {\n UI: {},\n Auth: {\n async getSession(c) {\n const sid = getCookie(c, \"axiom_session\");\n if (!sid) return null;\n const res = await sqlite.execute({\n sql: `SELECT * FROM ${TABLES.AGENTS} WHERE id = ?`,\n args: [sid],\n });\n return res.rows[0];\n },\n },\n };\n\n try {\n const logicRows = await sqlite.execute(\n `SELECT code FROM ${TABLES.KERNEL} WHERE name = 'main_logic'`,\n );\n const loader = new AsyncFunction(\n \"app\",\n \"scope\",\n \"sqlite\",\n \"TABLES\",\n \"IDENTITY\",\n \"DB\",\n \"AI\",\n \"getCookie\",\n \"setCookie\",\n \"html\",\n logicRows.rows[0][0],\n );\n await loader(\n runnerApp,\n scope,\n sqlite,\n TABLES,\n IDENTITY,\n DB,\n AI,\n getCookie,\n setCookie,\n html,\n );\n } catch (e) {\n return new Response(`KERNEL_ERROR: ${e.message}`, { status: 500 });\n }\n\n runnerApp.post(\"/create\", async (c) => {\n const { idea } = await c.req.parseBody();\n const rid = crypto.randomUUID();\n await sqlite.execute({\n sql:\n `INSERT INTO ${TABLES.REQUESTS} (id, status, description) VALUES (?, 'INIT', ?)`,\n args: [rid, idea as string],\n });\n return c.json({ id: rid });\n });\n\n runnerApp.get(\"/api/status/:id\", async (c) => {\n const res = await sqlite.execute({\n sql: `SELECT * FROM ${TABLES.REQUESTS} WHERE id=?`,\n args: [c.req.param(\"id\")],\n });\n const row = res.rows[0];\n if (!row) return c.json({ error: \"NOT_FOUND\" });\n if (row[5] === 1) {\n return c.json({ success: true, path: row[3], description: \"Stabilized\" });\n }\n\n if (row[1] === \"INIT\") {\n const meta = JSON.parse(\n await AI.query(\n `JSON ONLY: {\"name\":\"\", \"slug\":\"\", \"desc\":\"\", \"folder\":\"\", \"page\":\"\"}`,\n `Schema for: ${row[2]}`,\n ),\n );\n const slug = DB.formatSlug(meta.slug || meta.name);\n const path = `/s/${slug}/${meta.folder}/${meta.page}`;\n\n await sqlite.execute({\n sql:\n `INSERT OR IGNORE INTO ${TABLES.ROOTS} (name, slug, idea, description) VALUES (?, ?, ?, ?)`,\n args: [meta.name, slug, row[2], meta.desc],\n });\n await sqlite.execute({\n sql:\n `INSERT OR REPLACE INTO ${TABLES.NODES} (path, site_id, folder, title) VALUES (?, (SELECT id FROM ${TABLES.ROOTS} WHERE slug=?), ?, ?)`,\n args: [path, slug, meta.folder, meta.name],\n });\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.REQUESTS} SET status='BODY', description='WEAVING', path=? WHERE id=?`,\n args: [path, row[0]],\n });\n } else if (row[1] === \"BODY\") {\n const content = await AI.query(\n `HTML Body content only. Use Tailwind. Brutalist aesthetic.`,\n `Content for: ${row[2]}`,\n );\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.NODES} SET html = ?, generated = 1 WHERE path = ?`,\n args: [content, row[3]],\n });\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.REQUESTS} SET status='DONE', success=1 WHERE id=?`,\n args: [row[0]],\n });\n }\n return c.json({ success: false, description: row[2] });\n });\n\n return runnerApp.fetch(req);\n}",[],[256],{"_105":264,"_96":265,"_159":266,"_161":-5,"_162":144,"_163":267,"_165":166,"_107":268,"_101":269,"_169":170,"_171":270,"_173":22,"_174":-5,"_175":271,"_99":272},"73951536-f681-11f0-8b11-42dde27851f2","mew2.ts","9a3105ba-f645-11f0-942c-42dde27851f2",124,["D",1768969629878],27,"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\nimport { GoogleGenerativeAI } from \"npm:@google/generative-ai\";\nimport { html, raw } from \"npm:hono/html\";\n\nconst LOG_BUFFER_SIZE = 400;\nconst logs = [];\n\nfunction addLog(level, ...args) {\n const ts = Date.now();\n const msg = args.map((arg) =>\n typeof arg === \"object\" ? JSON.stringify(arg, null, 2) : String(arg)\n ).join(\" \");\n const entry = { ts, level, msg };\n logs.push(entry);\n if (logs.length > LOG_BUFFER_SIZE) logs.shift();\n try {\n // originalConsole[level](...args);\n } catch (_e) {\n // console.log(_e);\n }\n}\n\nconst originalConsole = {\n log: console.log,\n error: console.error,\n warn: console.warn,\n info: console.info,\n};\n\nconsole.log = (...args) => addLog(\"log\", ...args);\nconsole.error = (...args) => addLog(\"error\", ...args);\nconsole.warn = (...args) => addLog(\"warn\", ...args);\nconsole.info = (...args) => addLog(\"info\", ...args);\n\nconst originalFetch = globalThis.fetch;\nglobalThis.fetch = async (input, init = {}) => {\n const url = input instanceof Request ? input.url : String(input);\n const method = init.method || \"GET\";\n const start = performance.now();\n let status = 0, ok = false, bodyPreview = \"\";\n\n try {\n const res = await originalFetch(input, init);\n status = res.status;\n ok = res.ok;\n\n if (res.ok) {\n if (res.headers.get(\"content-type\")?.includes(\"json\")) {\n try {\n const clone = res.clone();\n const json = await clone.json();\n bodyPreview = JSON.stringify(json).slice(0, 200) +\n (JSON.stringify(json).length > 200 ? \"...\" : \"\");\n } catch {}\n } else {\n const clone = res.clone();\n bodyPreview = (await clone.text()).slice(0, 200) + \"...\";\n }\n }\n\n addLog(\n \"fetch\",\n `[${method} ${url}] ${status} ${ok ? \"OK\" : \"FAIL\"} ${\n Math.round(performance.now() - start)\n }ms` +\n (bodyPreview ? ` → ${bodyPreview}` : \"\"),\n );\n return res;\n } catch (err) {\n addLog(\"fetch\", `[${method} ${url}] ERROR ${err.message}`);\n throw err;\n }\n};\n\nconst SERVERS = [\n {\n id: \"google\",\n name: \"Gemini Nexus\",\n models: [\"gemini-1.5-flash-latest\", \"gemini-1.5-pro-latest\"],\n type: \"auth\",\n health: 1.0,\n },\n {\n id: \"pollinations\",\n name: \"Pollinations\",\n models: [],\n fallbackModels: [\n \"openai\",\n \"gemini-large\",\n \"claude\",\n \"deepseek\",\n \"qwen-coder\",\n \"mistral\",\n \"nomnom\",\n \"grok\",\n ],\n type: \"unauth\",\n base: \"https://gen.pollinations.ai/v1/chat/completions\",\n modelsUrl: \"https://gen.pollinations.ai/v1/models\",\n health: 0.8,\n },\n {\n id: \"huggingface\",\n name: \"HF Inference\",\n models: [\"meta-llama/Llama-3.3-70B-Instruct\"],\n type: \"unauth\",\n base: \"https://api-inference.huggingface.co/models/\",\n health: 0.9,\n },\n {\n id: \"puter\",\n name: \"Puter Cloud\",\n models: [\"gpt-5-nano\", \"claude-3.5-sonnet\"],\n type: \"unauth-get\",\n base: \"https://api.puter.com/v2/chat\",\n health: 0.6,\n },\n];\n\nconst GEMINI = new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\");\n\nconst HEALTH_DECAY = 0.4;\nconst HEALTH_RECOVER = 0.12;\nconst MIN_HEALTH_TO_TRY = 0.25;\n\nlet lastHealthTick = Date.now();\nconst backoff = new Map();\nconst modelCache = new Map();\nconst MODEL_CACHE_TTL = 15 * 60 * 1000;\n\nasync function loadModelsForServer(server) {\n if (server.models.length > 0 && !server.modelsUrl) {\n return server.models;\n }\n\n const cached = modelCache.get(server.id);\n const now = Date.now();\n if (cached && now - cached.fetchedAt < MODEL_CACHE_TTL) {\n return cached.models;\n }\n\n if (!server.modelsUrl) {\n return server.fallbackModels || server.models || [];\n }\n\n try {\n const res = await fetch(server.modelsUrl, {\n signal: AbortSignal.timeout(5000),\n });\n if (!res.ok) throw new Error(`models fetch ${res.status}`);\n\n const data = await res.json();\n let models = [];\n\n if (Array.isArray(data)) {\n models = data;\n } else if (data.data && Array.isArray(data.data)) {\n models = data.data.map((m) => m.id || m);\n } else if (Array.isArray(data.models)) {\n models = data.models;\n }\n\n models = models.filter(Boolean).map(String);\n\n if (models.length === 0) {\n models = server.fallbackModels || [];\n }\n\n modelCache.set(server.id, { models, fetchedAt: now });\n return models;\n } catch (err) {\n console.error(`Failed to load models for ${server.id}:`, err);\n const fb = server.fallbackModels || server.models || [];\n modelCache.set(server.id, { models: fb, fetchedAt: now });\n return fb;\n }\n}\n\nasync function initDynamicModels() {\n await Promise.allSettled(\n SERVERS.map(async (s) => {\n s.models = await loadModelsForServer(s);\n }),\n );\n console.log(\"[AXIOM] Model discovery complete\");\n}\n\nfunction isHighQuality(text) {\n if (!text || text.length < 80) return false;\n const lower = text.toLowerCase();\n if (\n lower.includes(\"rate limit\") || lower.includes(\"quota\") ||\n lower.includes(\"429\") || lower.includes(\"please try again later\") ||\n lower.length > 12000\n ) return false;\n\n const words = text.split(/\\s+/).length;\n if (words < 35) return false;\n\n if ((text.match(/(.{20,})\\1{2,}/g) || []).length > 1) return false;\n\n return true;\n}\n\nasync function dispatch(node, model, prompt, system, signal) {\n try {\n if (node.id === \"google\") {\n const m = GEMINI.getGenerativeModel({ model, systemInstruction: system });\n const res = await m.generateContent({\n contents: [{ role: \"user\", parts: [{ text: prompt }] }],\n });\n const txt = res.response.text();\n return isHighQuality(txt) ? txt : null;\n }\n\n if (true)// {node.type === \"unauth\" && node.base.includes(\"/v1/chat/completions\")) {\n {try{ const res = await fetch(node.base, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({\n model,\n messages: [\n ...(system ? [{ role: \"system\", content: system }] : []),\n { role: \"user\", content: prompt },\n ],\n temperature: 0.7,\n max_tokens: 2048,\n stream: false,\n }),\n signal,\n });\n\n if (!res.ok) throw new Error(`HTTP ${res.status}`);\n const data = await res.json();\n const txt = data.choices?.[0]?.message?.content ?? \"\";\n return isHighQuality(txt) ? txt : null;\n }catch(_e){}\n }\n\n if (node.type === \"unauth-get\") {\n const url = `${node.base}?model=${encodeURIComponent(model)}&prompt=${\n encodeURIComponent(prompt)\n }`;\n const res = await fetch(url, { signal });\n if (!res.ok) throw new Error(`HTTP ${res.status}`);\n const txt = await res.text();\n return isHighQuality(txt) ? txt : null;\n }\n\n if (true)//node.id === \"huggingface\") {\n {try{const url = `${node.base}${model}`;\n const res = await fetch(url, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({\n inputs: `${system ? system + \"\\n\\n\" : \"\"}${prompt}`,\n parameters: { max_new_tokens: 1024 },\n }),\n signal,\n });\n if (!res?.ok) throw new Error(`HTTP ${res?.status}`);\n const json = await res?.json();\n const txt = json[0]?.generated_text ?? \"\";\n return isHighQuality(txt) ? txt : null;\n }catch(_e){}\n }\n\n return null;\n } catch {\n return null;\n }\n}\n\nasync function callAI(prompt, system = \"\") {\n const now = Date.now();\n if (now - lastHealthTick > 60000) {\n SERVERS.forEach((s) => {\n s.health = Math.min(1, s.health + HEALTH_RECOVER);\n });\n lastHealthTick = now;\n }\n\n const tasks = SERVERS\n // .filter((s) => s.health >= MIN_HEALTH_TO_TRY)\n .flatMap((s) =>\n s.models.map((m, idx) => ({ node: s, model: m, priority: idx }))\n )\n .sort((a, b) => a.priority - b.priority || b.node.health - a.node.health);\n\n const controller = new AbortController();\n const timeoutId = setTimeout(() => controller.abort(), 11000);\n\n try {\n const winner = await Promise.any(\n tasks.map(async ({ node, model }) => {\n const delay = backoff.get(node.id) ?? 0;\n if (delay > now) await new Promise((r) => setTimeout(r, delay - now));\n\n const result = await dispatch(\n node,\n model,\n prompt,\n system,\n controller.signal,\n );\n\n if (result) {\n node.health = Math.min(1, node.health + 0.35);\n backoff.delete(node.id);\n controller.abort();\n return result;\n }\n\n node.health = Math.max(0.1, node.health * HEALTH_DECAY);\n const nextBackoff = Math.min(\n 60000,\n (backoff.get(node.id) ?? 1000) * 2.2 + 400,\n );\n backoff.set(node.id, now + nextBackoff);\nreturn null;\n // throw new Error(\"fail\");\n }),\n );\n\n return winner;\n } catch {\n return \"All providers failed or timed out. Try again in 30–60s.\";\n } finally {\n clearTimeout(timeoutId);\n }\n}\n\nconst TEST_PROMPTS = [\n {\n name: \"echo\",\n prompt: \"Respond exactly with: AXIOM-TEST-OK-2026\",\n expectSubstring: \"AXIOM-TEST-OK-2026\",\n minLength: 18,\n },\n];\n\nconst TEST_TIMEOUT_MS = 9000;\n\nasync function testModel(node, model) {\n const start = performance.now();\n const controller = new AbortController();\n const timeout = setTimeout(() => controller.abort(), TEST_TIMEOUT_MS);\n\n try {\n const response = await dispatch(\n node,\n model,\n TEST_PROMPTS[0].prompt,\n \"Strict test mode: be precise and short.\",\n controller.signal,\n );\n\n const duration = Math.round(performance.now() - start);\n\n if (!response) throw new Error(\"No response or failed quality gate\");\n\n const passed = response.includes(TEST_PROMPTS[0].expectSubstring) &&\n response.length >= TEST_PROMPTS[0].minLength &&\n !response.toLowerCase().includes(\"limit\") &&\n !response.toLowerCase().includes(\"error\");\n\n if (passed) {\n node.health = Math.min(1, node.health + 0.25);\n } else {\n node.health = Math.max(0.15, node.health * 0.65);\n }\n\n return {\n success: passed,\n latencyMs: duration,\n response: response.slice(0, 120) + (response.length > 120 ? \"...\" : \"\"),\n };\n } catch (err) {\n node.health = Math.max(0.1, node.health * 0.5);\n return {\n success: false,\n latencyMs: Math.round(performance.now() - start),\n error: err.message || \"Unknown error\",\n };\n } finally {\n clearTimeout(timeout);\n }\n}\n\nasync function runAllTests() {\n const results = {};\n\n await Promise.allSettled(\n SERVERS.map(async (node) => {\n results[node.id] = {};\n for (const model of node.models) {\n results[node.id][model] = await testModel(node, model);\n }\n }),\n );\n\n return results;\n}\n\nconst app = new Hono();\n\nsetTimeout(() => {\n Promise.allSettled([\n initDynamicModels(),\n runAllTests().then((r) => console.log(\"[AXIOM] Startup test results:\", r)),\n ]).catch(console.error);\n}, 1000);\n\napp.get(\"/\", (c) => {\n return c.html(`\n \n \n AXIOM MESH CONSOLE\n \n \n \n \n

AXIOM MESH

SYSTEM_READY // 2026.01.20

\n\n

\n \n ${\n raw(\n SERVERS.map((s) => `\n

${s.name}

\n ${s.type}\n

\n ${\n raw(\n s.models.map((m) => {\n // Escape for JS string: replace \\ → \\\\, ' → \\', \" → \\\"\n const safeModel = JSON.stringify(m).slice(1, -1); // JSON escapes properly\n return `\n \n `;\n }).join(\"\"),\n )\n }\n

\n `).join(\"\"),\n )\n }\n \n

\n\n

LIVE LOGS

\n \n

\n\n \n \n \n `);\n});\n\napp.get(\"/logs\", (c) => {\n const stream = new ReadableStream({\n start(controller) {\n const encoder = new TextEncoder();\n\n logs.forEach((log) => {\n controller.enqueue(encoder.encode(`data: ${JSON.stringify(log)}\\n\\n`));\n });\n\n const interval = setInterval(() => {\n controller.enqueue(encoder.encode(\": ping\\n\\n\"));\n }, 15000);\n\n c.req.signal.addEventListener(\"abort\", () => {\n clearInterval(interval);\n controller.close();\n });\n },\n });\n\n return new Response(stream, {\n headers: {\n \"Content-Type\": \"text/event-stream\",\n \"Cache-Control\": \"no-cache\",\n \"Connection\": \"keep-alive\",\n },\n });\n});\n// Prompt to generate 50 candlestick pattern groupings\nasync function generatePatternGroupings() {\n const prompt = `\n Generate 50 unique candlestick pattern groupings. \n Each grouping should have a short name suitable for a URL slug.\n Format as JSON: [{ \"name\": \"Grouping Name\", \"slug\": \"grouping-slug\" }]\n `;\n const raw = await callAI(prompt);\n try {\n return JSON.parse(raw);\n } catch {\n return [];\n }\n}\n\n// Prompt to generate patterns for a grouping\nasync function generatePatternsForGrouping(groupingName) {\n const prompt = `\n List all candlestick patterns in the grouping \"${groupingName}\".\n Provide JSON: [{ \"name\": \"Pattern Name\", \"slug\": \"pattern-slug\" }]\n `;\n const raw = await callAI(prompt);\n try {\n return JSON.parse(raw);\n } catch {\n return [];\n }\n}\n\n// Prompt to generate detailed pattern data\nasync function generatePatternDetails(patternName) {\n const prompt = `\n Provide detailed candlestick analysis for \"${patternName}\":\n - Exact body and wick measurements\n - Typical volume\n - Useful indicators\n - Socioeconomic context\n - Market type (bull, bear, range)\n - Price direction\n - Multi-timeframe interpretations\n - Take profit and stop loss\n - ASCII candle matrix (green/red with wicks on white background)\n Respond in JSON with keys: measurements, volume, indicators, context, market, direction, timeframe, risk, matrix\n `;\n const raw = await callAI(prompt);\n try {\n return JSON.parse(raw);\n } catch {\n return {};\n }\n}\n\n// Main grouping page\napp.get(\"/patterns\", async (c) => {\n const groupings = await generatePatternGroupings();\n return c.html(`\n \n Candlestick Pattern Groupings\n \n

Candlestick Pattern Groupings

${g?.name}

\n \n \n `);\n});\n\n// Grouping page with pattern links\napp.get(\"/patterns/:group\", async (c) => {\n const { group } = c.req.param();\n const patterns = await generatePatternsForGrouping(group);\n return c.html(`\n \n Patterns in ${group}\n \n

Patterns in ${group}

${p.name}

\n \n \n `);\n});\n\n// Individual pattern page with full details\napp.get(\"/patterns/:group/:pattern\", async (c) => {\n const { group, pattern } = c.req.param();\n const details = await generatePatternDetails(pattern);\n return c.html(`\n \n ${pattern}\n \n

${pattern}

Measurements

${\n    JSON.stringify(details.measurements, null, 2)\n  }

Volume

${details.volume}

Indicators

${details.indicators}

Socioeconomic Context

${details.context}

Market / Direction

${details.market} / ${details.direction}

Timeframes

${\n    JSON.stringify(details.timeframe, null, 2)\n  }

Risk (Take Profit / Stop Loss)

${\n    JSON.stringify(details.risk, null, 2)\n  }

Pattern Matrix

${details.matrix}

\n \n \n `);\n});\n\napp.get(\"/test\", async (c) => {\n console.log(`test`);\n const { server, model } = c.req.query();\n console.log(`test ${server},${model}`);\n const node = SERVERS.find((s) => s.id === server);\n if (!node) {\n console.log(`test !node`);\n return c.json({ success: false });\n }\n const ok = await dispatch(\n node,\n model,\n \"Respond exactly OK\",\n \"Strict\",\n new AbortController().signal,\n );\n\n return c.json({ success: ok === \"OK\" });\n});\n\napp.get(\"/tests\", async (c) => {\n const results = await runAllTests();\n return c.json(results);\n});\n\napp.get(\"/health\", (c) => {\n const summary = SERVERS.map((s) => ({\n id: s.id,\n health: s.health.toFixed(2),\n models: s.models.map((m) => ({\n model: m,\n status: s.health > 0.4 ? \"healthy\" : s.health > 0.2 ? \"degraded\" : \"down\",\n })),\n }));\n return c.json({ status: \"ok\", providers: summary });\n});\n\napp.post(\"/chat\", async (c) => {\n const { prompt, system = \"\" } = await c.req.json();\n if (!prompt) return c.json({ error: \"prompt required\" }, 400);\n\n const result = await callAI(prompt, system);\n return c.text(result);\n});\n\nexport default app.fetch;",[],[265],{"_105":274,"_96":275,"_159":276,"_161":-5,"_162":144,"_163":277,"_165":166,"_107":278,"_101":279,"_169":170,"_171":280,"_173":22,"_174":-5,"_175":281,"_99":282},"db428ab0-f63b-11f0-a827-42dde27851f2","new-file-1812.ts","c1b35df4-f63b-11f0-bda7-42dde27851f2",11,["D",1768939739041],2,"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\nimport { html, raw } from \"npm:hono/html\";\nimport { sqlite } from \"https://esm.town/v/std/sqlite\";\nimport { getCookie, setCookie } from \"npm:hono/cookie\";\nimport { GoogleGenerativeAI } from \"npm:@google/generative-ai\";\n\n// --- SYSTEM CONSTANTS ---\nconst VERSION = 145;\nconst DB_PREFIX = `AXIOM_V${VERSION}`;\n\nconst TABLES = {\n AGENTS: `${DB_PREFIX}_AGENTS`,\n ROOTS: `${DB_PREFIX}_ROOTS`,\n NODES: `${DB_PREFIX}_NODES`,\n FLOW: `${DB_PREFIX}_FLOW`,\n KERNEL: `${DB_PREFIX}_KERNEL`, // Stores source code for self-reflection\n REQUESTS: `${DB_PREFIX}_VECTORS`,\n META: `${DB_PREFIX}_META`,\n};\n\nconst IDENTITY = {\n version: `${VERSION}.0.0-AXIOM`,\n owner: \"Michael Morton\",\n year: \"2026\",\n};\n\n// --- THE KERNEL (THE LOGIC) ---\n// We define this as a named exportable function.\n// This allows us to run it directly AND serialize it.\nconst KERNEL = async (CTX) => {\n const { app, scope, sqlite, TABLES, IDENTITY, html, raw } = CTX;\n\n // 1. UI COMPONENT SYSTEM\n scope.UI.Layout = ({ children, title, user, siteCss }) => {\n return html`\n \n \n \n \n \n ${title} | AXIOM\n \n \n \n \n \n AXIOMATIC\n

\n Directory\n \n ${user ? user.username : \"VOID_IDENTITY\"}\n

\n \n \n

\n ${children}\n

\n \n PROTOCOL_v${IDENTITY.version}\n ${IDENTITY.owner} // ${IDENTITY.year}\n \n \n \n `;\n };\n\n // 2. ROUTING LOGIC\n app.get(\"/\", async (c) => {\n const user = await scope.Auth.getSession(c);\n // Fetch recent flows (sites) created by users\n const flowRes = await sqlite.execute(\n `SELECT name, idea FROM ${TABLES.FLOW} ORDER BY ts DESC LIMIT 8`,\n );\n\n const fragments = flowRes.rows.map((f) =>\n html`\n \n \n ${f[0]}\n \n

\n ${f[1]}\n

\n \n `\n );\n\n return c.html(scope.UI.Layout({\n title: \"Nexus\",\n user,\n children: html`\n

\n \n

\n \n \n AWAITING_INPUT\n

\n \n System_Ready\n

\n \n Manifest\n \n \n \n \n\n

\n Recent_Flows\n

${fragments}

\n\n \n \n `,\n }));\n });\n\n app.get(\"/s\", async (c) => {\n const user = await scope.Auth.getSession(c);\n const res = await sqlite.execute(\n `SELECT name, slug, description FROM ${TABLES.ROOTS} ORDER BY ts DESC`,\n );\n const list = res.rows.map((s) =>\n html`\n \n

\n \n ${s[0]}\n

\n ${s[2]}\n

\n \n \n →\n \n \n `\n );\n\n return c.html(scope.UI.Layout({\n title: \"Directory\",\n user,\n children: html`\n

The_Archive

${list}

\n `,\n }));\n });\n\n // Dynamic Site Router\n app.get(\"/s/:slug/:folder/:page\", async (c) => {\n const user = await scope.Auth.getSession(c);\n const path = c.req.path;\n const slug = c.req.param(\"slug\");\n\n const siteRes = await sqlite.execute({\n sql: `SELECT css FROM ${TABLES.ROOTS} WHERE slug = ?`,\n args: [slug],\n });\n\n const nodeRes = await sqlite.execute({\n sql:\n `SELECT html, title FROM ${TABLES.NODES} WHERE path = ? AND generated = 1`,\n args: [path],\n });\n\n if (!nodeRes.rows[0]) return c.redirect(\"/\");\n\n return c.html(scope.UI.Layout({\n title: nodeRes.rows[0][1],\n user,\n siteCss: siteRes.rows[0]?.[0],\n children: html`\n

\n \n HOME /\n ${slug\n .toUpperCase()} /\n ${c.req.param(\"folder\")}\n \n

\n ${raw(nodeRes.rows[0][0])}\n

\n `,\n }));\n });\n};\n\n// --- DATABASE & TOOLS ---\nconst DB = {\n async init() {\n // 1. Meta Table\n await sqlite.execute(\n `CREATE TABLE IF NOT EXISTS ${TABLES.META} (key TEXT PRIMARY KEY, value TEXT)`,\n );\n\n // 2. Version Check\n let currentVer = null;\n try {\n const res = await sqlite.execute(\n `SELECT value FROM ${TABLES.META} WHERE key = 'version'`,\n );\n currentVer = res.rows[0]?.[0];\n } catch (e) {}\n\n // 3. Schema Migration\n if (currentVer !== IDENTITY.version) {\n console.log(\n `[SYSTEM] Migrating from ${currentVer} to ${IDENTITY.version}`,\n );\n for (const t of Object.values(TABLES)) {\n if (t !== TABLES.META) {\n try {\n await sqlite.execute(`DROP TABLE IF EXISTS ${t}`);\n } catch (e) {}\n }\n }\n await sqlite.execute(`DELETE FROM ${TABLES.META} WHERE key='version'`);\n await sqlite.execute({\n sql: `INSERT INTO ${TABLES.META} (key, value) VALUES ('version', ?)`,\n args: [IDENTITY.version],\n });\n }\n\n // 4. Create Tables\n const schema = [\n `CREATE TABLE IF NOT EXISTS ${TABLES.AGENTS} (id INTEGER PRIMARY KEY, username TEXT UNIQUE, email TEXT UNIQUE, pass TEXT, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.ROOTS} (id INTEGER PRIMARY KEY, name TEXT, slug TEXT UNIQUE, idea TEXT, description TEXT, css TEXT, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.NODES} (path TEXT PRIMARY KEY, site_id INTEGER, folder TEXT, title TEXT, html TEXT, description TEXT, generated INTEGER DEFAULT 0, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.FLOW} (id INTEGER PRIMARY KEY, name TEXT, idea TEXT, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.KERNEL} (name TEXT PRIMARY KEY, code TEXT)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.REQUESTS} (id TEXT PRIMARY KEY, status TEXT, description TEXT, path TEXT, success INTEGER DEFAULT 0, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n ];\n\n for (const sql of schema) await sqlite.execute(sql);\n\n // 5. Seed Initial Data\n const qCount = await sqlite.execute(`SELECT COUNT(*) FROM ${TABLES.FLOW}`);\n if (qCount.rows[0][0] < 1) {\n await sqlite.execute({\n sql: `INSERT INTO ${TABLES.FLOW} (name, idea) VALUES (?, ?)`,\n args: [\"Recursive Logic\", \"Systems that rewrite their own core.\"],\n });\n }\n\n // 6. Backup Kernel Source (Self-Reflection)\n await this.seedKernel();\n },\n\n async seedKernel() {\n // We store the source code in the DB just for \"reflection\" purposes,\n // ensuring the AI knows its own source code if queried.\n const sourceCode = KERNEL.toString();\n await sqlite.execute({\n sql: `INSERT OR REPLACE INTO ${TABLES.KERNEL} (name, code) VALUES (?, ?)`,\n args: [\"main_logic\", sourceCode],\n });\n },\n\n formatSlug: (s) =>\n (s || \"void\").toLowerCase().trim().replace(/[^\\w\\s-]/g, \"\").replace(\n /[\\s_-]+/g,\n \"-\",\n ).replace(/^-+|-+$/g, \"\"),\n};\n\nconst AI = {\n genAI: new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\"),\n async query(system, prompt) {\n try {\n const model = this.genAI.getGenerativeModel({\n model: \"gemini-1.5-flash\",\n systemInstruction: system,\n });\n const result = await model.generateContent(prompt);\n const text = result.response.text();\n return text.replace(/```(json|html|javascript|css)?/g, \"\").replace(\n /```/g,\n \"\",\n ).trim();\n } catch (e) {\n console.error(\"AI_FAULT\", e);\n return null;\n }\n },\n};\n\n// --- ENTRY POINT ---\nexport default async function (req) {\n // 1. Initialize Infrastructure\n await DB.init();\n const app = new Hono();\n\n // 2. Define Scope\n const scope = {\n UI: {},\n Auth: {\n async getSession(c) {\n const sid = getCookie(c, \"axiom_session\");\n if (!sid) return null;\n const res = await sqlite.execute({\n sql: `SELECT * FROM ${TABLES.AGENTS} WHERE id = ?`,\n args: [sid],\n });\n const row = res.rows[0];\n return row ? { id: row[0], username: row[1] } : null;\n },\n },\n };\n\n // 3. Dependency Injection Context\n const CTX = { app, scope, sqlite, TABLES, IDENTITY, DB, AI, html, raw };\n\n // 4. EXECUTE KERNEL\n // We execute the kernel directly rather than rehydrating it from string.\n // This guarantees scope access and prevents the \"Kernel Factory\" crash.\n try {\n await KERNEL(CTX);\n } catch (e) {\n return new Response(`KERNEL PANIC: ${e.message}\\n${e.stack}`, {\n status: 500,\n });\n }\n\n // --- ASYNC API HANDLERS (Background Workers) ---\n app.post(\"/create\", async (c) => {\n const body = await c.req.parseBody();\n const idea = body.idea;\n const rid = crypto.randomUUID();\n await sqlite.execute({\n sql:\n `INSERT INTO ${TABLES.REQUESTS} (id, status, description) VALUES (?, 'INIT', ?)`,\n args: [rid, idea],\n });\n return c.json({ id: rid });\n });\n\n app.get(\"/api/status/:id\", async (c) => {\n const rid = c.req.param(\"id\");\n const res = await sqlite.execute({\n sql: `SELECT * FROM ${TABLES.REQUESTS} WHERE id=?`,\n args: [rid],\n });\n const row = res.rows[0];\n\n if (!row) return c.json({ error: \"NOT_FOUND\" });\n if (row[4] === 1) {\n return c.json({ success: true, path: row[3], description: \"Stabilized\" });\n }\n\n // AI Generation Logic\n if (row[1] === \"INIT\") {\n const resp = await AI.query(\n `Return valid JSON ONLY. Structure: {\"name\": \"string\", \"slug\": \"string\", \"desc\": \"string\", \"folder\": \"string\", \"page\": \"string\"}`,\n `Create a web architecture for: ${row[2]}`,\n );\n if (!resp) return c.json({ description: \"NEURAL_FAILURE\" });\n\n try {\n const meta = JSON.parse(resp);\n const slug = DB.formatSlug(meta.slug || meta.name);\n const path = `/s/${slug}/${meta.folder}/${meta.page}`;\n\n await sqlite.execute({\n sql:\n `INSERT OR IGNORE INTO ${TABLES.ROOTS} (name, slug, idea, description) VALUES (?, ?, ?, ?)`,\n args: [meta.name, slug, row[2], meta.desc],\n });\n await sqlite.execute({\n sql:\n `INSERT OR REPLACE INTO ${TABLES.NODES} (path, site_id, folder, title) VALUES (?, (SELECT id FROM ${TABLES.ROOTS} WHERE slug=?), ?, ?)`,\n args: [path, slug, meta.folder, meta.name],\n });\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.REQUESTS} SET status='BODY', description='WEAVING', path=? WHERE id=?`,\n args: [path, row[0]],\n });\n // Add to recent flows immediately so it shows up in the list\n await sqlite.execute({\n sql:\n `INSERT OR IGNORE INTO ${TABLES.FLOW} (name, idea) VALUES (?, ?)`,\n args: [meta.name, row[2]],\n });\n } catch (e) {\n return c.json({ description: \"JSON_PARSE_ERROR\" });\n }\n } else if (row[1] === \"BODY\") {\n const content = await AI.query(\n `Return raw HTML content only. Use Tailwind classes. Dark brutalist aesthetics. NO or tags. Just the inner content.`,\n `Generate content for: ${row[2]}`,\n );\n if (!content) return c.json({ description: \"GENERATION_FAILURE\" });\n\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.NODES} SET html = ?, generated = 1 WHERE path = ?`,\n args: [content, row[3]],\n });\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.REQUESTS} SET status='DONE', success=1 WHERE id=?`,\n args: [row[0]],\n });\n }\n\n return c.json({ success: false, description: row[1] });\n });\n\n return app.fetch(req);\n}",[],[275],{"_105":284,"_96":285,"_159":286,"_161":-5,"_162":144,"_163":287,"_165":166,"_107":288,"_101":168,"_169":170,"_171":289,"_173":22,"_174":-5,"_175":290,"_99":291},"07a79f6a-f645-11f0-84af-42dde27851f2","new-file-340.ts","62688a74-f643-11f0-8f09-42dde27851f2",24,["D",1768943678996],"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\nimport { GoogleGenerativeAI } from \"npm:@google/generative-ai\";\nimport { html, raw } from \"npm:hono/html\";\n\n// ────────────────────────────────────────────────────────────────────────\n// PROVIDERS + HEALTH + DYNAMIC MODELS\n// ────────────────────────────────────────────────────────────────────────\n\nconst SERVERS = [\n {\n id: \"google\",\n name: \"Gemini Nexus\",\n models: [\"gemini-1.5-flash-latest\", \"gemini-1.5-pro-latest\"],\n type: \"auth\",\n health: 1.0,\n },\n {\n id: \"pollinations\",\n name: \"Pollinations\",\n models: [], // ← populated dynamically\n fallbackModels: [\n \"openai\",\n \"gemini-large\",\n \"claude\",\n \"deepseek\",\n \"qwen-coder\",\n \"mistral\",\n \"nomnom\",\n \"grok\",\n ],\n type: \"unauth\",\n base: \"https://gen.pollinations.ai/v1/chat/completions\",\n modelsUrl: \"https://gen.pollinations.ai/v1/models\",\n health: 0.8,\n },\n {\n id: \"huggingface\",\n name: \"HF Inference\",\n models: [\"meta-llama/Llama-3.3-70B-Instruct\"],\n type: \"unauth\",\n base: \"https://api-inference.huggingface.co/models/\",\n health: 0.9,\n },\n {\n id: \"puter\",\n name: \"Puter Cloud\",\n models: [\"gpt-5-nano\", \"claude-3.5-sonnet\"],\n type: \"unauth-get\",\n base: \"https://api.puter.com/v2/chat\",\n health: 0.6,\n },\n];\n\nconst GEMINI = new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\");\n\nconst HEALTH_DECAY = 0.4;\nconst HEALTH_RECOVER = 0.12;\nconst MIN_HEALTH_TO_TRY = 0.25;\n\nlet lastHealthTick = Date.now();\nconst backoff = new Map();\n\nconst modelCache = new Map();\nconst MODEL_CACHE_TTL = 15 * 60 * 1000; // 15 minutes\n\n// ────────────────────────────────────────────────────────────────────────\n// DYNAMIC MODEL LOADER\n// ────────────────────────────────────────────────────────────────────────\n\nasync function loadModelsForServer(server) {\n if (server.models.length > 0 && !server.modelsUrl) {\n return server.models;\n }\n\n const cached = modelCache.get(server.id);\n const now = Date.now();\n if (cached && now - cached.fetchedAt < MODEL_CACHE_TTL) {\n return cached.models;\n }\n\n if (!server.modelsUrl) {\n return server.fallbackModels || server.models || [];\n }\n\n try {\n const res = await fetch(server.modelsUrl, {\n signal: AbortSignal.timeout(5000),\n });\n if (!res.ok) throw new Error(`models fetch ${res.status}`);\n\n const data = await res.json();\n let models = [];\n\n if (Array.isArray(data)) {\n models = data;\n } else if (data.data && Array.isArray(data.data)) {\n models = data.data.map((m) => m.id || m);\n } else if (Array.isArray(data.models)) {\n models = data.models;\n }\n\n models = models.filter(Boolean).map(String);\n\n if (models.length === 0) {\n models = server.fallbackModels || [];\n }\n\n modelCache.set(server.id, { models, fetchedAt: now });\n return models;\n } catch (err) {\n console.error(`Failed to load models for ${server.id}:`, err);\n const fb = server.fallbackModels || server.models || [];\n modelCache.set(server.id, { models: fb, fetchedAt: now });\n return fb;\n }\n}\n\nasync function initDynamicModels() {\n await Promise.allSettled(\n SERVERS.map(async (s) => {\n s.models = await loadModelsForServer(s);\n }),\n );\n console.log(\"[AXIOM] Model discovery complete\");\n}\n\n// Start model discovery (non-blocking)\nsetTimeout(initDynamicModels, 500);\n\n// ────────────────────────────────────────────────────────────────────────\n// QUALITY GATE\n// ────────────────────────────────────────────────────────────────────────\n\nfunction isHighQuality(text) {\n if (!text || text.length < 80) return false;\n const lower = text.toLowerCase();\n if (\n lower.includes(\"rate limit\") || lower.includes(\"quota\") ||\n lower.includes(\"429\") || lower.includes(\"please try again later\") ||\n lower.length > 12000\n ) return false;\n\n const words = text.split(/\\s+/).length;\n if (words < 35) return false;\n\n if ((text.match(/(.{20,})\\1{2,}/g) || []).length > 1) return false;\n\n return true;\n}\n\n// ────────────────────────────────────────────────────────────────────────\n// DISPATCH\n// ────────────────────────────────────────────────────────────────────────\n\nasync function dispatch(node, model, prompt, system, signal) {\n try {\n if (node.id === \"google\") {\n const m = GEMINI.getGenerativeModel({ model, systemInstruction: system });\n const res = await m.generateContent({\n contents: [{ role: \"user\", parts: [{ text: prompt }] }],\n signal,\n });\n const txt = res.response.text();\n return isHighQuality(txt) ? txt : null;\n }\n\n if (node.type === \"unauth\" && node.base.includes(\"/v1/chat/completions\")) {\n const res = await fetch(node.base, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({\n model,\n messages: [\n ...(system ? [{ role: \"system\", content: system }] : []),\n { role: \"user\", content: prompt },\n ],\n temperature: 0.7,\n max_tokens: 2048,\n stream: false,\n }),\n signal,\n });\n\n if (!res.ok) throw new Error(`HTTP ${res.status}`);\n const data = await res.json();\n const txt = data.choices?.[0]?.message?.content ?? \"\";\n return isHighQuality(txt) ? txt : null;\n }\n\n if (node.type === \"unauth-get\") {\n const url = `${node.base}?model=${encodeURIComponent(model)}&prompt=${\n encodeURIComponent(prompt)\n }`;\n const res = await fetch(url, { signal });\n if (!res.ok) throw new Error(`HTTP ${res.status}`);\n const txt = await res.text();\n return isHighQuality(txt) ? txt : null;\n }\n\n if (node.id === \"huggingface\") {\n const url = `${node.base}${model}`;\n const res = await fetch(url, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify({\n inputs: `${system ? system + \"\\n\\n\" : \"\"}${prompt}`,\n parameters: { max_new_tokens: 1024 },\n }),\n signal,\n });\n if (!res.ok) throw new Error(`HTTP ${res.status}`);\n const json = await res.json();\n const txt = json[0]?.generated_text ?? \"\";\n return isHighQuality(txt) ? txt : null;\n }\n\n return null;\n } catch {\n return null;\n }\n}\n\n// ────────────────────────────────────────────────────────────────────────\n// RACE + HEALTH + BACKOFF\n// ────────────────────────────────────────────────────────────────────────\n\nasync function callAI(prompt, system = \"\") {\n const now = Date.now();\n if (now - lastHealthTick > 60000) {\n SERVERS.forEach((s) => {\n s.health = Math.min(1, s.health + HEALTH_RECOVER);\n });\n lastHealthTick = now;\n }\n\n const tasks = SERVERS\n .filter((s) => s.health >= MIN_HEALTH_TO_TRY)\n .flatMap((s) =>\n s.models.map((m, idx) => ({ node: s, model: m, priority: idx }))\n )\n .sort((a, b) => a.priority - b.priority || b.node.health - a.node.health);\n\n const controller = new AbortController();\n const timeoutId = setTimeout(() => controller.abort(), 11000);\n\n try {\n const winner = await Promise.any(\n tasks.map(async ({ node, model }) => {\n const delay = backoff.get(node.id) ?? 0;\n if (delay > now) await new Promise((r) => setTimeout(r, delay - now));\n\n const result = await dispatch(\n node,\n model,\n prompt,\n system,\n controller.signal,\n );\n\n if (result) {\n node.health = Math.min(1, node.health + 0.35);\n backoff.delete(node.id);\n controller.abort();\n return result;\n }\n\n node.health = Math.max(0.1, node.health * HEALTH_DECAY);\n const nextBackoff = Math.min(\n 60000,\n (backoff.get(node.id) ?? 1000) * 2.2 + 400,\n );\n backoff.set(node.id, now + nextBackoff);\n\n throw new Error(\"fail\");\n }),\n );\n\n return winner;\n } catch {\n return \"All providers failed or timed out. Try again in 30–60s.\";\n } finally {\n clearTimeout(timeoutId);\n }\n}\n\n// ────────────────────────────────────────────────────────────────────────\n// TEST SUITE\n// ────────────────────────────────────────────────────────────────────────\n\nconst TEST_PROMPTS = [\n {\n name: \"echo\",\n prompt: \"Respond exactly with: AXIOM-TEST-OK-2026\",\n expectSubstring: \"AXIOM-TEST-OK-2026\",\n minLength: 18,\n },\n];\n\nconst TEST_TIMEOUT_MS = 9000;\n\nasync function testModel(node, model) {\n const start = performance.now();\n const controller = new AbortController();\n const timeout = setTimeout(() => controller.abort(), TEST_TIMEOUT_MS);\n\n try {\n const response = await dispatch(\n node,\n model,\n TEST_PROMPTS[0].prompt,\n \"Strict test mode: be precise and short.\",\n controller.signal,\n );\n\n const duration = Math.round(performance.now() - start);\n\n if (!response) throw new Error(\"No response or failed quality gate\");\n\n const passed = response.includes(TEST_PROMPTS[0].expectSubstring) &&\n response.length >= TEST_PROMPTS[0].minLength &&\n !response.toLowerCase().includes(\"limit\") &&\n !response.toLowerCase().includes(\"error\");\n\n if (passed) {\n node.health = Math.min(1, node.health + 0.25);\n } else {\n node.health = Math.max(0.15, node.health * 0.65);\n }\n\n return {\n success: passed,\n latencyMs: duration,\n response: response.slice(0, 120) + (response.length > 120 ? \"...\" : \"\"),\n };\n } catch (err) {\n node.health = Math.max(0.1, node.health * 0.5);\n return {\n success: false,\n latencyMs: Math.round(performance.now() - start),\n error: err.message || \"Unknown error\",\n };\n } finally {\n clearTimeout(timeout);\n }\n}\n\nasync function runAllTests() {\n const results = {};\n\n await Promise.allSettled(\n SERVERS.map(async (node) => {\n results[node.id] = {};\n for (const model of node.models) {\n results[node.id][model] = await testModel(node, model);\n }\n }),\n );\n\n return results;\n}\n\n// ────────────────────────────────────────────────────────────────────────\n// APP\n// ────────────────────────────────────────────────────────────────────────\n\nconst app = new Hono();\n\n// Startup tests & model discovery\nsetTimeout(() => {\n Promise.allSettled([\n initDynamicModels(),\n runAllTests().then((r) => console.log(\"[AXIOM] Startup test results:\", r)),\n ]).catch(console.error);\n}, 1000);\n\napp.get(\"/\", (c) => {\n return c.html(\n `\n \n AXIOM MESH CONSOLE\n \n \n \n \n

AXIOM MESH

SYSTEM_READY // 2026.01.20

\n\n

\n ${\n raw(\n SERVERS.map((s) => `\n

${s.name}

\n ${s.type}\n

\n ${\n s.models.map((m) => `\n \n `).join(\"\")\n }\n

\n `).join(\"\"),\n )\n }\n

\n\n \n \n `,\n );\n});\n\napp.get(\"/test\", async (c) => {\n const { server, model } = c.req.query();\n const node = SERVERS.find((s) => s.id === server);\n if (!node) return c.json({ success: false });\n\n const ok = await dispatch(\n node,\n model,\n \"Respond exactly OK\",\n \"Strict\",\n new AbortController().signal,\n );\n\n return c.json({ success: ok === \"OK\" });\n});\n\napp.get(\"/tests\", async (c) => {\n const results = await runAllTests();\n return c.json(results);\n});\n\napp.get(\"/health\", (c) => {\n const summary = SERVERS.map((s) => ({\n id: s.id,\n health: s.health.toFixed(2),\n models: s.models.map((m) => ({\n model: m,\n status: s.health > 0.4 ? \"healthy\" : s.health > 0.2 ? \"degraded\" : \"down\",\n })),\n }));\n return c.json({ status: \"ok\", providers: summary });\n});\n\napp.post(\"/chat\", async (c) => {\n const { prompt, system = \"\" } = await c.req.json();\n if (!prompt) return c.json({ error: \"prompt required\" }, 400);\n\n const result = await callAI(prompt, system);\n return c.text(result);\n});\n\nexport default app.fetch;",[],[285],{"_105":293,"_96":294,"_159":295,"_161":-5,"_162":144,"_163":296,"_165":166,"_107":297,"_101":193,"_169":170,"_171":298,"_173":22,"_174":-5,"_175":299,"_99":300},"9133fee2-f64d-11f0-bd9b-42dde27851f2","new-file-5208.ts","65d8bf44-f64d-11f0-a1c3-42dde27851f2",52,["D",1768947345740],"import { env, pipeline } from \"jsr:@vergedx/transformers\";\n\n// Force WebGPU backend\nenv.backends.onnx.wasm.wasmPaths =\n \"https://cdn.jsdelivr.net/npm/onnxruntime-web/dist/\";\nenv.allowLocalModels = false;\nenv.useBrowserCache = true;\n\nasync function testLlama3() {\n console.log(\"Loading Llama-3.2-3B-Instruct...\");\n const generator = await pipeline(\n \"text-generation\",\n \"Xenova/llama-3.2-3b-instruct\",\n );\n\n console.log(\"Running inference...\");\n const out = await generator(\"Explain quantum entanglement simply:\", {\n max_new_tokens: 80,\n temperature: 0.7,\n });\n\n console.log(\"Output:\", out[0].generated_text);\n\n return {\n model: \"llama-3.2-3b-instruct\",\n ok: out[0].generated_text.length > 20,\n text: out[0].generated_text,\n };\n}\n\nexport default testLlama3;",[],[294],{"_105":302,"_96":303,"_159":304,"_161":-5,"_162":144,"_163":305,"_165":166,"_107":306,"_101":249,"_169":170,"_171":307,"_173":22,"_174":-5,"_175":308,"_99":309},"71e4fe44-f641-11f0-9844-42dde27851f2","new-file-5281.ts","fba59970-f63e-11f0-be30-42dde27851f2",18,["D",1768942139222],"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\nimport { GoogleGenerativeAI } from \"npm:@google/generative-ai\";\nimport { sqlite } from \"https://esm.town/v/std/sqlite\";\nimport { html, raw } from \"npm:hono/html\";\n// --- PROVIDER REGISTRY ---\nconst SERVERS = [\n {\n id: \"google\",\n name: \"Gemini Nexus\",\n models: [\"gemini-2.0-flash\", \"gemini-1.5-pro\"],\n type: \"auth\",\n },\n {\n id: \"pollinations\",\n name: \"Pollinations Core\",\n models: [\"deepseek-r1\", \"qwen3-coder\", \"mistral-large\"],\n type: \"unauth\",\n base: \"https://text.pollinations.ai\",\n },\n {\n id: \"puter\",\n name: \"Puter Cloud\",\n models: [\"gpt-4.1-nano\", \"claude-3.7-sonnet\"],\n type: \"unauth\",\n base: \"https://api.puter.com/v2/chat\",\n },\n {\n id: \"huggingface\",\n name: \"HF Serverless\",\n models: [\"meta-llama/Llama-3.3-70B-Instruct\"],\n type: \"unauth\",\n base: \"https://api-inference.huggingface.co/models/\",\n },\n {\n id: \"shadow\",\n name: \"Shadow Mesh\",\n models: [\"llama-3.1-8b\"],\n type: \"unauth\",\n base: \"https://text.shadow.io\",\n }, // 2026 Spore Clone\n];\n\nconst AI_KERNEL = {\n genAI: new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\"),\n\n /**\n * Universal Dispatcher: Handles routing to any provider type\n */\n async dispatch(node, model, prompt, system) {\n const controller = new AbortController();\n const timeout = setTimeout(() => controller.abort(), 12000); // 12s Hard limit\n\n try {\n if (node.type === \"auth\" && node.id === \"google\") {\n const m = this.genAI.getGenerativeModel({\n model,\n systemInstruction: system,\n });\n const res = await m.generateContent(prompt);\n return res.response.text();\n }\n\n // REST Generic (Pollinations/Clones Style)\n let url = `${node.base}/${\n encodeURIComponent(prompt)\n }?model=${model}&system=${encodeURIComponent(system)}`;\n\n // Provider-specific adjustments\n if (node.id === \"puter\") {\n url = `${node.base}?model=${model}&prompt=${\n encodeURIComponent(prompt)\n }`;\n }\n if (node.id === \"huggingface\") url = `${node.base}${model}`;\n\n const res = await fetch(url, { signal: controller.signal });\n if (!res.ok) throw new Error(`HTTP_${res.status}`);\n const text = await res.text();\n\n // Content-Aware Validation (Check for Rate Limit strings in 200 OK)\n if (text.toLowerCase().includes(\"rate limit\") || text.length < 2) {\n throw new Error(\"FAKE_OK\");\n }\n return text;\n } finally {\n clearTimeout(timeout);\n }\n },\n\n /**\n * Racing Pattern: Hits top 3 nodes, first valid return wins\n */\n async CallAI(prompt, system) {\n const racePool = SERVERS.flatMap((s) =>\n s.models.slice(0, 1).map((m) => ({ node: s, model: m }))\n );\n return Promise.any(racePool.map(async (task) => {\n const result = await this.dispatch(task.node, task.model, prompt, system);\n if (result) return result;\n throw new Error(\"Empty\");\n }));\n },\n};\n\nconst app = new Hono();\n\n// --- DASHBOARD UI ---\napp.get(\"/\", (c) => {\n return c.html(\n `\n \n AI MESH CONSOLE\n \n \n \n \n

AXIOM MESH

SYSTEM_READY // 2026.01.20

\n\n

\n ${\n raw(\n SERVERS.map((s) => (`\n

${s.name}

\n ${s.type}\n

\n ${\n raw(\n s.models.map((m) => (`\n \n `)).join(\"\"),\n )\n }\n

\n `)).join(\"\"),\n )\n }\n

\n\n \n \n `,\n );\n});\n\n// --- API LAYER ---\napp.get(\"/test\", async (c) => {\n const { server, model } = c.req.query();\n const node = SERVERS.find((s) => s.id === server);\n const success = await AI_KERNEL.dispatch(\n node,\n model,\n \"Respond with OK\",\n \"Strict Mode\",\n ).then((t) => t?.includes(\"OK\"));\n return c.json({ success });\n});\n\nexport default app.fetch;",[],[303],{"_105":311,"_96":312,"_159":313,"_161":-5,"_162":144,"_163":314,"_165":166,"_107":315,"_101":32,"_169":316,"_171":73,"_173":22,"_174":-5,"_175":317,"_99":318},"5b5e5e3e-f643-11f0-8a53-42dde27851f2","new-file-5410.ts","5b5e5f1a-f643-11f0-8a53-42dde27851f2",19,["D",1768942960456],"script",[],[312],{"_105":320,"_96":321,"_159":322,"_161":-5,"_162":144,"_163":102,"_165":22,"_107":323,"_101":258,"_169":170,"_171":324,"_173":22,"_174":-5,"_175":325,"_99":326},"f9490d2e-f7ba-11f0-910f-42dde27851f2","new-file-7150.tsx","19151b30-f7ba-11f0-931a-42dde27851f2",["D",1769104286667],"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\n\nconst app = new Hono();\nconst PROMPT = \"Explain entropy in 5 words.\";\n\napp.get(\"/\", (c) => {\n return c.html(\n \n \n POLLINATIONS_FUNCTIONAL_V9\n \n \n \n

// CLASSIFIED_FUNCTIONAL_NODES_V9

\n \n

\n\n JSON.stringify({messages: [{role:\"user\", content: p}], model: \"openai\"}),\n OAI_NO_MODEL: (p) => JSON.stringify({messages: [{role:\"user\", content: p}]}),\n RAW_TEXT: (p) => p,\n FORM_DATA: (p) => \"prompt=\" + encodeURIComponent(p),\n NONE: () => null\n };\n\n function init() {\n const root = document.getElementById('matrix-root');\n let html = '';\n FUNCTIONAL_GROUPS.forEach(group => {\n group.configs.forEach((c, i) => {\n const id = \\`\\${group.tag}-\\${i}\\`;\n html += \\`\n

\n \\${group.name}\n

\\${c.m} | \\${c.ep.replace('https://','')}

IDLE...

\\`;\n });\n });\n root.innerHTML = html;\n }\n\n async function runFunctionalMatrix() {\n for (const group of FUNCTIONAL_GROUPS) {\n group.configs.forEach(async (c, i) => {\n const id = \\`\\${group.tag}-\\${i}\\`;\n const resBox = document.getElementById(\"res-\" + id);\n resBox.innerText = \"PROBING...\";\n\n try {\n const r = await fetch(c.ep, {\n method: c.m,\n headers: c.v === \"RAW_TEXT\" ? {\"Content-Type\":\"text/plain\"} : {\"Content-Type\":\"application/json\"},\n body: Payloads[c.v](P)\n });\n const raw = await r.text();\n let display = raw;\n let thought = \"\";\n\n try {\n const json = JSON.parse(raw);\n display = json.choices?.[0]?.message?.content || json.content || raw;\n if(json.choices?.[0]?.message?.reasoning_content) {\n thought = json.choices[0].message.reasoning_content;\n }\n } catch(e) {}\n\n resBox.innerHTML = \\`

\\${display}

\\` + \n (thought ? \\`REASONING: \\${thought.substring(0,150)}...\\` : \"\");\n } catch(err) {\n resBox.innerText = \"OFFLINE: \" + err.message;\n }\n });\n }\n }\n\n init();\n `,\n }}\n />\n \n ,\n );\n});\n\nexport default app.fetch;",[],[321],{"_105":328,"_96":329,"_159":330,"_161":-5,"_162":144,"_163":331,"_165":166,"_107":332,"_101":279,"_169":170,"_171":333,"_173":22,"_174":-5,"_175":334,"_99":335},"69705084-f63b-11f0-bd22-42dde27851f2","new-file-7433.ts","5f2da81a-f63b-11f0-80fa-42dde27851f2",8,["D",1768939548088],"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\nimport { html, raw } from \"npm:hono/html\";\nimport { sqlite } from \"https://esm.town/v/std/sqlite\";\nimport { getCookie, setCookie } from \"npm:hono/cookie\";\nimport { GoogleGenerativeAI } from \"npm:@google/generative-ai\";\n\n// --- CONFIGURATION ---\nconst VERSION = 145; // Increment this to force a DB schema refresh\nconst DB_PREFIX = `AXIOM_V${VERSION}`;\n\nconst TABLES = {\n AGENTS: `${DB_PREFIX}_AGENTS`,\n ROOTS: `${DB_PREFIX}_ROOTS`,\n NODES: `${DB_PREFIX}_NODES`,\n FLOW: `${DB_PREFIX}_FLOW`,\n KERNEL: `${DB_PREFIX}_KERNEL`,\n REQUESTS: `${DB_PREFIX}_VECTORS`,\n META: `${DB_PREFIX}_META`,\n};\n\nconst IDENTITY = {\n version: `${VERSION}.0.0-AXIOM`,\n owner: \"Michael Morton\",\n year: \"2026\",\n};\n\n// --- THE KERNEL (PURE LOGIC) ---\n// This function is defined natively. We serialize it to the DB.\n// No backtick escaping is required here.\nconst KERNEL_LOGIC = async (CTX) => {\n const { app, scope, sqlite, TABLES, IDENTITY, html, raw } = CTX;\n\n // 1. UI COMPONENT SYSTEM\n scope.UI.Layout = ({ children, title, user, siteCss }) => {\n return html`\n \n \n \n \n \n ${title} | AXIOM\n \n \n \n \n \n AXIOMATIC\n

\n Directory\n \n ${user ? user.username : \"VOID_IDENTITY\"}\n

\n \n \n

\n ${children}\n

\n \n PROTOCOL_v${IDENTITY.version}\n ${IDENTITY.owner} // ${IDENTITY.year}\n \n \n \n `;\n };\n\n // 2. ROUTING LOGIC\n app.get(\"/\", async (c) => {\n const user = await scope.Auth.getSession(c);\n const flowRes = await sqlite.execute(\n `SELECT name, idea FROM ${TABLES.FLOW} ORDER BY ts DESC LIMIT 8`,\n );\n\n // We render fragments using Hono's html helper\n const fragments = flowRes.rows.map((f) =>\n html`\n \n \n ${f[0]}\n \n

${f[1]}

\n \n `\n );\n\n return c.html(scope.UI.Layout({\n title: \"Nexus\",\n user,\n children: html`\n

\n \n

\n \n \n AWAITING_INPUT\n

\n \n System_Ready\n

\n \n Manifest\n \n \n \n \n\n

\n Recent_Flows\n

${fragments}

\n \n ${s[0]}\n

\n ${s[2]}\n

\n \n \n →\n \n \n `\n );\n\n return c.html(scope.UI.Layout({\n title: \"Directory\",\n user,\n children: html`\n

The_Archive

${list}

\n `,\n }));\n });\n\n app.get(\"/s/:slug/:folder/:page\", async (c) => {\n const user = await scope.Auth.getSession(c);\n const path = c.req.path;\n const slug = c.req.param(\"slug\");\n\n const siteRes = await sqlite.execute({\n sql: `SELECT css FROM ${TABLES.ROOTS} WHERE slug = ?`,\n args: [slug],\n });\n\n const nodeRes = await sqlite.execute({\n sql:\n `SELECT html, title FROM ${TABLES.NODES} WHERE path = ? AND generated = 1`,\n args: [path],\n });\n\n if (!nodeRes.rows[0]) return c.redirect(\"/\");\n\n // Inject RAW HTML securely\n const content = raw(nodeRes.rows[0][0]);\n\n return c.html(scope.UI.Layout({\n title: nodeRes.rows[0][1],\n user,\n siteCss: siteRes.rows[0]?.[0],\n children: html`\n

\n \n HOME /\n ${slug\n .toUpperCase()} /\n ${c.req.param(\"folder\")}\n \n

\n ${content}\n

\n `,\n }));\n });\n};\n\n// --- SYSTEM TOOLS ---\nconst DB = {\n async init() {\n await sqlite.execute(\n `CREATE TABLE IF NOT EXISTS ${TABLES.META} (key TEXT PRIMARY KEY, value TEXT)`,\n );\n\n // Self-Healing Schema Check\n let currentVer = null;\n try {\n const res = await sqlite.execute(\n `SELECT value FROM ${TABLES.META} WHERE key = 'version'`,\n );\n currentVer = res.rows[0]?.[0];\n } catch (e) {}\n\n if (currentVer !== IDENTITY.version) {\n console.log(\n `[SYSTEM] Migrating from ${currentVer} to ${IDENTITY.version}`,\n );\n // Drop old tables to ensure clean state on version change\n for (const t of Object.values(TABLES)) {\n if (t !== TABLES.META) {\n try {\n await sqlite.execute(`DROP TABLE IF EXISTS ${t}`);\n } catch (e) {}\n }\n }\n await sqlite.execute(`DELETE FROM ${TABLES.META} WHERE key='version'`);\n await sqlite.execute({\n sql: `INSERT INTO ${TABLES.META} (key, value) VALUES ('version', ?)`,\n args: [IDENTITY.version],\n });\n }\n\n // Initialize Schema\n const schema = [\n `CREATE TABLE IF NOT EXISTS ${TABLES.AGENTS} (id INTEGER PRIMARY KEY, username TEXT UNIQUE, email TEXT UNIQUE, pass TEXT, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.ROOTS} (id INTEGER PRIMARY KEY, name TEXT, slug TEXT UNIQUE, idea TEXT, description TEXT, css TEXT, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.NODES} (path TEXT PRIMARY KEY, site_id INTEGER, folder TEXT, title TEXT, html TEXT, description TEXT, generated INTEGER DEFAULT 0, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.FLOW} (id INTEGER PRIMARY KEY, name TEXT, idea TEXT, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.KERNEL} (name TEXT PRIMARY KEY, code TEXT)`,\n `CREATE TABLE IF NOT EXISTS ${TABLES.REQUESTS} (id TEXT PRIMARY KEY, status TEXT, description TEXT, path TEXT, success INTEGER DEFAULT 0, ts DATETIME DEFAULT CURRENT_TIMESTAMP)`,\n ];\n\n for (const sql of schema) await sqlite.execute(sql);\n\n // Seed Initial Data\n const qCount = await sqlite.execute(`SELECT COUNT(*) FROM ${TABLES.FLOW}`);\n if (qCount.rows[0][0] < 1) {\n await sqlite.execute({\n sql: `INSERT INTO ${TABLES.FLOW} (name, idea) VALUES (?, ?)`,\n args: [\"Recursive Logic\", \"Systems that rewrite their own core.\"],\n });\n }\n\n await this.seedKernel();\n },\n\n async seedKernel() {\n // SYMBOLIC SERIALIZATION: Function -> String\n const sourceCode = KERNEL_LOGIC.toString();\n await sqlite.execute({\n sql: `INSERT OR REPLACE INTO ${TABLES.KERNEL} (name, code) VALUES (?, ?)`,\n args: [\"main_logic\", sourceCode],\n });\n },\n\n formatSlug: (s) =>\n (s || \"void\").toLowerCase().trim().replace(/[^\\w\\s-]/g, \"\").replace(\n /[\\s_-]+/g,\n \"-\",\n ).replace(/^-+|-+$/g, \"\"),\n};\n\nconst AI = {\n genAI: new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\"),\n async query(system, prompt) {\n try {\n const model = this.genAI.getGenerativeModel({\n model: \"gemini-1.5-flash\",\n systemInstruction: system,\n });\n const result = await model.generateContent(prompt);\n const text = result.response.text();\n // Sanitize markdown code blocks\n return text.replace(/```(json|html|javascript|css)?/g, \"\").replace(\n /```/g,\n \"\",\n ).trim();\n } catch (e) {\n console.error(\"AI_FAULT\", e);\n return null;\n }\n },\n};\n\n// --- ENTRY POINT ---\nexport default async function (req) {\n await DB.init();\n const app = new Hono();\n\n // Define Scope\n const scope = {\n UI: {},\n Auth: {\n async getSession(c) {\n const sid = getCookie(c, \"axiom_session\");\n if (!sid) return null;\n const res = await sqlite.execute({\n sql: `SELECT * FROM ${TABLES.AGENTS} WHERE id = ?`,\n args: [sid],\n });\n const row = res.rows[0];\n return row ? { id: row[0], username: row[1] } : null;\n },\n },\n };\n\n try {\n // 1. Fetch Logic from DB\n const logicRows = await sqlite.execute(\n `SELECT code FROM ${TABLES.KERNEL} WHERE name = 'main_logic'`,\n );\n if (!logicRows.rows[0]) throw new Error(\"KERNEL_MISSING\");\n\n // 2. Rehydrate Function\n // We construct a factory that returns the logic function\n const AsyncFunction =\n Object.getPrototypeOf(async function () {}).constructor;\n const kernelFactory = new AsyncFunction(\n `return (${logicRows.rows[0][0]})`,\n )();\n\n // 3. Inject Context\n const CTX = { app, scope, sqlite, TABLES, IDENTITY, DB, AI, html, raw };\n await kernelFactory(CTX);\n } catch (e) {\n return new Response(`KERNEL PANIC: ${e.message}\\n${e.stack}`, {\n status: 500,\n });\n }\n\n // --- MINING API (External to Kernel for Stability) ---\n app.post(\"/create\", async (c) => {\n const body = await c.req.parseBody();\n const idea = body.idea;\n const rid = crypto.randomUUID();\n await sqlite.execute({\n sql:\n `INSERT INTO ${TABLES.REQUESTS} (id, status, description) VALUES (?, 'INIT', ?)`,\n args: [rid, idea],\n });\n return c.json({ id: rid });\n });\n\n app.get(\"/api/status/:id\", async (c) => {\n const rid = c.req.param(\"id\");\n const res = await sqlite.execute({\n sql: `SELECT * FROM ${TABLES.REQUESTS} WHERE id=?`,\n args: [rid],\n });\n const row = res.rows[0]; // [id, status, description, path, success, ts]\n\n if (!row) return c.json({ error: \"NOT_FOUND\" });\n if (row[4] === 1) {\n return c.json({ success: true, path: row[3], description: \"Stabilized\" });\n }\n\n // AI PIPELINE\n if (row[1] === \"INIT\") {\n const resp = await AI.query(\n `Return valid JSON ONLY. No markdown. Structure: {\"name\": \"string\", \"slug\": \"string\", \"desc\": \"string\", \"folder\": \"string\", \"page\": \"string\"}`,\n `Create a web architecture for: ${row[2]}`,\n );\n if (!resp) return c.json({ description: \"NEURAL_FAILURE\" });\n\n try {\n const meta = JSON.parse(resp);\n const slug = DB.formatSlug(meta.slug || meta.name);\n const path = `/s/${slug}/${meta.folder}/${meta.page}`;\n\n await sqlite.execute({\n sql:\n `INSERT OR IGNORE INTO ${TABLES.ROOTS} (name, slug, idea, description) VALUES (?, ?, ?, ?)`,\n args: [meta.name, slug, row[2], meta.desc],\n });\n await sqlite.execute({\n sql:\n `INSERT OR REPLACE INTO ${TABLES.NODES} (path, site_id, folder, title) VALUES (?, (SELECT id FROM ${TABLES.ROOTS} WHERE slug=?), ?, ?)`,\n args: [path, slug, meta.folder, meta.name],\n });\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.REQUESTS} SET status='BODY', description='WEAVING', path=? WHERE id=?`,\n args: [path, row[0]],\n });\n } catch (e) {\n return c.json({ description: \"JSON_PARSE_ERROR\" });\n }\n } else if (row[1] === \"BODY\") {\n const content = await AI.query(\n `Return raw HTML content only. Use Tailwind classes. Dark brutalist aesthetics. NO or tags. Just the inner content.`,\n `Generate content for: ${row[2]}`,\n );\n if (!content) return c.json({ description: \"GENERATION_FAILURE\" });\n\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.NODES} SET html = ?, generated = 1 WHERE path = ?`,\n args: [content, row[3]],\n });\n await sqlite.execute({\n sql:\n `UPDATE ${TABLES.REQUESTS} SET status='DONE', success=1 WHERE id=?`,\n args: [row[0]],\n });\n }\n\n return c.json({ success: false, description: row[1] });\n });\n\n return app.fetch(req);\n}",[],[329],{"_105":337,"_96":338,"_159":339,"_161":-5,"_162":144,"_163":269,"_165":166,"_107":340,"_101":279,"_169":170,"_171":341,"_173":22,"_174":-5,"_175":342,"_99":343},"54a6a02c-f645-11f0-ab0f-42dde27851f2","new-file-7900.ts","4f38d86c-f645-11f0-9677-42dde27851f2",["D",1768943808178],"/** @jsxImportSource npm:hono@3/jsx */\nimport { Hono } from \"npm:hono\";\nimport { GoogleGenerativeAI } from \"npm:@google/generative-ai\";\nimport { html, raw } from \"npm:hono/html\";\n\n// ────────────────────────────────────────────────────────────────────────\n// LOG BUFFER + SSE BROADCAST\n// ────────────────────────────────────────────────────────────────────────\n\nconst LOG_BUFFER_SIZE = 400;\nconst logs = []; // { ts: number, level: 'log'|'error'|'warn'|'info'|'fetch', msg: string }\n\nfunction addLog(level, ...args) {\n const ts = Date.now();\n const msg = args.map(arg =>\n typeof arg === \"object\" ? JSON.stringify(arg, null, 2) : String(arg)\n ).join(\" \");\n const entry = { ts, level, msg };\n logs.push(entry);\n if (logs.length > LOG_BUFFER_SIZE) logs.shift();\n\n // Also keep original console output\n originalConsole[level](...args);\n}\n\nconst originalConsole = {\n log: console.log,\n error: console.error,\n warn: console.warn,\n info: console.info,\n};\n\nconsole.log = (...args) => addLog(\"log\", ...args);\nconsole.error = (...args) => addLog(\"error\", ...args);\nconsole.warn = (...args) => addLog(\"warn\", ...args);\nconsole.info = (...args) => addLog(\"info\", ...args);\n\n// ────────────────────────────────────────────────────────────────────────\n// FETCH WRAPPER WITH LOGGING\n// ────────────────────────────────────────────────────────────────────────\n\nconst originalFetch = globalThis.fetch;\nglobalThis.fetch = async (input, init = {}) => {\n const url = input instanceof Request ? input.url : String(input);\n const method = init.method || \"GET\";\n const start = performance.now();\n\n let status = 0, ok = false, bodyPreview = \"\";\n\n try {\n const res = await originalFetch(input, init);\n status = res.status;\n ok = res.ok;\n\n if (res.ok && res.headers.get(\"content-type\")?.includes(\"json\")) {\n try {\n const clone = res.clone();\n const json = await clone.json();\n bodyPreview = JSON.stringify(json).slice(0, 200) + (JSON.stringify(json).length > 200 ? \"...\" : \"\");\n } catch {}\n } else if (res.ok) {\n const clone = res.clone();\n bodyPreview = (await clone.text()).slice(0, 200) + \"...\";\n }\n\n addLog(\"fetch\", `[${method} ${url}] ${status} ${ok ? \"OK\" : \"FAIL\"} ${Math.round(performance.now() - start)}ms` +\n (bodyPreview ? ` → ${bodyPreview}` : \"\"));\n return res;\n } catch (err) {\n addLog(\"fetch\", `[${method} ${url}] ERROR ${err.message}`);\n throw err;\n }\n};\n\n// ────────────────────────────────────────────────────────────────────────\n// PROVIDERS + HEALTH + DYNAMIC MODELS (unchanged except logging already captured)\n// ────────────────────────────────────────────────────────────────────────\n\nconst SERVERS = [\n { id: \"google\", name: \"Gemini Nexus\", models: [\"gemini-1.5-flash-latest\", \"gemini-1.5-pro-latest\"], type: \"auth\", health: 1.0 },\n { id: \"pollinations\", name: \"Pollinations\", models: [], fallbackModels: [\"openai\",\"gemini-large\",\"claude\",\"deepseek\",\"qwen-coder\",\"mistral\",\"nomnom\",\"grok\"], type: \"unauth\", base: \"https://gen.pollinations.ai/v1/chat/completions\", modelsUrl: \"https://gen.pollinations.ai/v1/models\", health: 0.8 },\n { id: \"huggingface\", name: \"HF Inference\", models: [\"meta-llama/Llama-3.3-70B-Instruct\"], type: \"unauth\", base: \"https://api-inference.huggingface.co/models/\", health: 0.9 },\n { id: \"puter\", name: \"Puter Cloud\", models: [\"gpt-5-nano\", \"claude-3.5-sonnet\"], type: \"unauth-get\", base: \"https://api.puter.com/v2/chat\", health: 0.6 },\n];\n\nconst GEMINI = new GoogleGenerativeAI(Deno.env.get(\"GEMINI_API_KEY\") || \"\");\n\nconst HEALTH_DECAY = 0.4;\nconst HEALTH_RECOVER = 0.12;\nconst MIN_HEALTH_TO_TRY = 0.25;\n\nlet lastHealthTick = Date.now();\nconst backoff = new Map();\nconst modelCache = new Map();\nconst MODEL_CACHE_TTL = 15 * 60 * 1000;\n\n// Dynamic model loader, test suite, dispatch, callAI remain mostly unchanged\n// (fetch logging wrapper already captures modelUrl fetches, /chat calls, etc.)\n\n// ... (keep loadModelsForServer, initDynamicModels, isHighQuality, dispatch, callAI, testModel, runAllTests as before)\n\n// ────────────────────────────────────────────────────────────────────────\n// APP + LOGGING UI\n// ────────────────────────────────────────────────────────────────────────\n\nconst app = new Hono();\n\n// Startup\nsetTimeout(() => {\n Promise.allSettled([\n initDynamicModels(),\n runAllTests().then(r => console.log(\"[AXIOM] Startup test results:\", r))\n ]).catch(console.error);\n}, 1000);\n\n// Dashboard with log panel\napp.get(\"/\", (c) => {\n return c.html(`\n \n \n AXIOM MESH CONSOLE\n \n \n \n \n

AXIOM MESH

SYSTEM_READY // 2026.01.20

\n\n \n

\n ${raw(SERVERS.map(s => /* ... your provider HTML ... */).join(\"\"))}\n

\n\n \n

LIVE LOGS

\n \n

\n\n \n\n \n \n \n \n `);\n});\n\n// SSE logs endpoint\napp.get(\"/logs\", (c) => {\n const stream = new ReadableStream({\n start(controller) {\n const encoder = new TextEncoder();\n\n // Send recent logs on connect\n logs.forEach(log => {\n controller.enqueue(encoder.encode(`data: ${JSON.stringify(log)}\\n\\n`));\n });\n\n // Keep-alive + future logs\n const interval = setInterval(() => {\n controller.enqueue(encoder.encode(\": ping\\n\\n\"));\n }, 15000);\n\n // Cleanup on disconnect\n c.req.signal.addEventListener(\"abort\", () => {\n clearInterval(interval);\n controller.close();\n });\n }\n });\n\n return new Response(stream, {\n headers: {\n \"Content-Type\": \"text/event-stream\",\n \"Cache-Control\": \"no-cache\",\n \"Connection\": \"keep-alive\"\n }\n });\n});\n\n// Keep your other routes: /test, /tests, /health, /chat\n\nexport default app.fetch;",[],[338],{"_105":345,"_96":346,"_159":347,"_161":-5,"_162":144,"_163":348,"_165":166,"_107":349,"_101":350,"_169":170,"_171":351,"_173":22,"_174":-5,"_175":352,"_99":353},"701d2e60-f65f-11f0-9b17-42dde27851f2","new-file-9058.ts","e752850a-f64d-11f0-bafc-42dde27851f2",112,["D",1768955021164],59,"export default function () {\n return new Response(\n `\n\n\n \n WebGPU LLM Test\n\n\n

Loading…

\n\n\n\n`,\n { headers: { \"Content-Type\": \"text/html\" } },\n );\n}\n/*Starting…\nDiagnostics:[object Object]\nCreating pipeline…\nError: SyntaxError: Unexpected token '<', \"\n typeof arg === \"object\" ? JSON.stringify(arg, null, 2) : String(arg)\n )\n .join(\" \");\n const entry = { ts, level, msg };\n logs.push(entry);\n if (logs.length > LOG_BUFFER_SIZE) logs.shift();\n try {\n originalConsole[level](...args);\n } catch (_e) {\n console.log(_e);\n }\n}\n\nconst originalConsole = {\n log: console.log,\n error: console.error,\n warn: console.warn,\n info: console.info,\n};\n\nconsole.log = (...args) => addLog(\"log\", ...args);\nconsole.error = (...args) => addLog(\"error\", ...args);\nconsole.warn = (...args) => addLog(\"warn\", ...args);\nconsole.info = (...args) => addLog(\"info\", ...args);\n\nconst originalFetch = globalThis.fetch;\nglobalThis.fetch = async (input, init = {}) => {\n const url = input instanceof Request ? input.url : String(input);\n const method = init.method || \"GET\";\n const start = performance.now();\n let status = 0,\n ok = false,\n bodyPreview = \"\";\n\n try {\n const res = await originalFetch(input, init);\n status = res.status;\n ok = res.ok;\n\n if (res.ok) {\n if (res.headers.get(\"content-type\")?.includes(\"json\")) {\n try {\n const clone = res.clone();\n const json = await clone.json();\n const s = JSON.stringify(json);\n bodyPreview = s.slice(0, 200) + (s.length > 200 ? \"...\" : \"\");\n } catch {}\n } else {\n const clone = res.clone();\n bodyPreview = (await clone.text()).slice(0, 200) + \"...\";\n }\n }\n\n addLog(\n \"fetch\",\n `[${method} ${url}] ${status} ${ok ? \"OK\" : \"FAIL\"} ${\n Math.round(performance.now() - start)\n }ms` + (bodyPreview ? ` → ${bodyPreview}` : \"\"),\n );\n return res;\n } catch (err) {\n addLog(\"fetch\", `[${method} ${url}] ERROR ${err.message}`);\n throw err;\n }\n};\n\n/* ------------------------- HF LOCAL-ONLY CONFIG ------------------------- */\n\nconst HF_BASE = \"https://api-inference.huggingface.co/models/\";\nconst HF_DEFAULT_MODEL = \"meta-llama/Llama-3.3-70B-Instruct\";\n\nconst HEALTH_DECAY = 0.4;\nconst HEALTH_RECOVER = 0.12;\nconst MIN_HEALTH_TO_TRY = 0.25;\n\nlet lastHealthTick = Date.now();\nlet hfHealth = 0.9;\nconst backoff = new Map();\nconst modelCache = new Map();\nconst MODEL_CACHE_TTL = 15 * 60 * 1000;\n\n/* ------------------------- CAPABILITIES + TASK INFERENCE ------------------------- */\n\nasync function hfDiscoverCapabilities(model) {\n const key = `hf-cap-${model}`;\n const cached = modelCache.get(key);\n const now = Date.now();\n if (cached && now - cached.fetchedAt < MODEL_CACHE_TTL) return cached.cap;\n\n const url = `${HF_BASE}${encodeURIComponent(model)}`;\n const res = await fetch(url, { method: \"GET\" });\n\n let cap = {\n supportsTextGen: true,\n supportsChat: true,\n maxTokens: 2048,\n };\n\n if (res.ok) {\n const meta = await res.json();\n const tags = meta.tags || [];\n cap.supportsTextGen = tags.includes(\"text-generation\") ||\n tags.includes(\"causal-lm\") ||\n cap.supportsTextGen;\n cap.supportsChat = tags.includes(\"chat\") ||\n tags.includes(\"instruction-tuned\") ||\n cap.supportsChat;\n cap.maxTokens = meta.max_length || cap.maxTokens;\n }\n\n modelCache.set(key, { cap, fetchedAt: now });\n return cap;\n}\n\nfunction inferTaskType(prompt, system) {\n const text = `${system} ${prompt}`.toLowerCase();\n if (text.includes(\"summarize\") || text.includes(\"tl;dr\")) {\n return \"summarization\";\n }\n if (text.includes(\"translate\")) return \"translation\";\n if (text.includes(\"classify\")) return \"classification\";\n return \"generation\";\n}\n\nfunction estimateTokens(text) {\n return Math.ceil(text.split(/\\s+/).length * 1.4);\n}\n\nfunction ensureTokenBudget(prompt, system, maxTokens) {\n const total = estimateTokens(prompt) + estimateTokens(system);\n if (total > maxTokens * 0.9) {\n return prompt.slice(0, Math.floor(prompt.length * 0.6));\n }\n return prompt;\n}\n\nfunction isHighQuality(text) {\n if (!text || text.length < 80) return false;\n const lower = text.toLowerCase();\n if (\n lower.includes(\"rate limit\") ||\n lower.includes(\"quota\") ||\n lower.includes(\"429\") ||\n lower.includes(\"please try again later\") ||\n lower.length > 12000\n ) {\n return false;\n }\n\n const words = text.split(/\\s+/).length;\n if (words < 35) return false;\n\n if ((text.match(/(.{20,})\\1{2,}/g) || []).length > 1) return false;\n\n return true;\n}\n\n/* ------------------------- HF ERROR NORMALIZATION ------------------------- */\n\nfunction normalizeHFError(status, body) {\n if (status === 503) return \"model-loading\";\n if (status === 429) return \"rate-limit\";\n if (status >= 500) return \"server-error\";\n\n const msg = body?.error || \"\";\n if (msg.includes(\"Input too long\")) return \"input-too-long\";\n if (msg.includes(\"Model\")) return \"model-error\";\n\n return \"unknown\";\n}\n\n/* ------------------------- CORE HF DISPATCH ------------------------- */\n\nasync function dispatchHF(model, prompt, system, signal) {\n const cap = await hfDiscoverCapabilities(model);\n const task = inferTaskType(prompt, system); // currently informational; could tune params per task\n\n prompt = ensureTokenBudget(prompt, system, cap.maxTokens || 2048);\n\n const url = `${HF_BASE}${encodeURIComponent(model)}`;\n const input = `${system ? system + \"\\n\\n\" : \"\"}${prompt}`;\n\n const body = {\n inputs: input,\n parameters: {\n max_new_tokens: cap.maxTokens || 2048,\n temperature: 0.7,\n return_full_text: false,\n },\n options: {\n wait_for_model: false,\n use_cache: true,\n },\n };\n\n const res = await fetch(url, {\n method: \"POST\",\n headers: { \"Content-Type\": \"application/json\" },\n body: JSON.stringify(body),\n signal,\n });\n\n const status = res.status;\n let json = null;\n try {\n json = await res.json();\n } catch {}\n\n if (!res.ok) {\n const type = normalizeHFError(status, json);\n console.warn(\"HF error:\", type, json);\n return null;\n }\n\n let txt = json?.generated_text ||\n json?.[0]?.generated_text ||\n json?.[0]?.summary_text ||\n \"\";\n\n if (!txt && typeof json === \"string\") txt = json;\n\n return isHighQuality(txt) ? txt : null;\n}\n\n/* ------------------------- BEST-EFFORT LOCAL-ONLY CALL ------------------------- */\n\nasync function hfBestEffort(prompt, system = \"\") {\n const now = Date.now();\n\n if (now - lastHealthTick > 60000) {\n hfHealth = Math.min(1, hfHealth + HEALTH_RECOVER);\n lastHealthTick = now;\n }\n\n if (hfHealth < MIN_HEALTH_TO_TRY) {\n return \"Provider temporarily degraded. Try again in 30–60 seconds.\";\n }\n\n const controller = new AbortController();\n const timeoutId = setTimeout(() => controller.abort(), 11000);\n\n const delay = backoff.get(\"hf\") ?? 0;\n if (delay > now) {\n await new Promise((r) => setTimeout(r, delay - now));\n }\n\n try {\n const primary = await dispatchHF(\n HF_DEFAULT_MODEL,\n prompt,\n system,\n controller.signal,\n );\n\n if (primary) {\n hfHealth = Math.min(1, hfHealth + 0.35);\n backoff.delete(\"hf\");\n return primary;\n }\n\n hfHealth = Math.max(0.1, hfHealth * HEALTH_DECAY);\n const nextBackoff = Math.min(\n 60000,\n (backoff.get(\"hf\") ?? 1000) * 2.2 + 400,\n );\n backoff.set(\"hf\", now + nextBackoff);\n\n // Speculative retry with same model, shorter timeout\n const controller2 = new AbortController();\n const timeoutId2 = setTimeout(() => controller2.abort(), 8000);\n\n try {\n const alt = await dispatchHF(\n HF_DEFAULT_MODEL,\n prompt,\n system,\n controller2.signal,\n );\n if (alt) {\n hfHealth = Math.min(1, hfHealth + 0.25);\n backoff.delete(\"hf\");\n return alt;\n }\n } finally {\n clearTimeout(timeoutId2);\n }\n\n return \"Generation failed. Try again shortly.\";\n } catch {\n hfHealth = Math.max(0.1, hfHealth * HEALTH_DECAY);\n return \"Provider error or timeout. Try again shortly.\";\n } finally {\n clearTimeout(timeoutId);\n }\n}\n\n/* ------------------------- HONO APP WIRED TO LOCAL-ONLY HF ------------------------- */\n\nconst app = new Hono();\n\napp.get(\"/\", (c) => {\n return c.html(\n html`\n \n \n \n Local-Only HF Orchestrator\n \n \n

Local-Only HF Orchestrator

\n \n

Recent Logs

${raw(JSON.stringify(logs.slice(-40), null, 2))}

\n \n \n `,\n );\n});\n\napp.post(\"/chat\", async (c) => {\n const body = await c.req.parseBody();\n const prompt = String(body.prompt || \"\");\n const system =\n \"You are an elite, precise assistant. Respond with deeply reasoned, high-signal, non-repetitive answers.\";\n\n const answer = await hfBestEffort(prompt, system);\n return c.json({ answer });\n});\n\nexport default app.fetch;",[],[356],"files",[156,178,187,197,206,216,225,233,243,254,263,273,283,292,301,310,319,327,336,344,354],"counts",{"_367":32,"_368":369,"_370":32,"_371":372,"_375":376},"pulls","branches",1,"forks","envVars",{"_373":279,"_374":32},"count","emptyCount","connections",{"_373":32,"_377":32},"unfilledCount","contents",[],"file",{"_105":207,"_96":100,"_159":208,"_161":-5,"_162":144,"_163":209,"_165":166,"_107":382,"_101":211,"_169":170,"_171":212,"_173":22,"_174":213,"_99":383},["D",1768979574902],[100],"upstreamProject","upstreamBranch","emptyEnvVars","isNew","flags",[],"routes/_app.x.$handle.$name.branch.$branch.version.$version.code.$",{"_94":95,"_96":97,"_98":44,"_101":102,"_99":100,"_103":392,"_154":397,"_363":461,"_380":462,"_378":464,"_465":-5,"_466":-5,"_467":-5,"_468":-5,"_469":470,"_371":-5},{"_105":106,"_96":97,"_107":393,"_109":-5,"_110":22,"_111":112,"_113":-5,"_114":-5,"_14":115,"_116":-5,"_117":22,"_118":-5,"_119":-5,"_120":32,"_16":121,"_122":-5,"_123":-5,"_124":125,"_126":125,"_127":128,"_129":128,"_130":128,"_131":125,"_132":125,"_23":394,"_136":395,"_141":396,"_153":22},["D",1768938649513],{"_94":95,"_134":135},{"_94":95,"_138":135,"_14":115,"_139":140},{"_101":143,"_96":44,"_105":144,"_145":106,"_146":-5,"_147":-5,"_148":-5,"_149":-5,"_150":-5,"_151":97,"_152":95},[398,401,404,407,410,413,416,419,422,425,428,431,434,437,440,443,446,449,452,455,458],{"_105":157,"_96":158,"_159":160,"_161":-5,"_162":144,"_163":164,"_165":166,"_107":167,"_101":168,"_169":170,"_171":172,"_173":22,"_174":-5,"_175":399,"_99":400},[],[158],{"_105":179,"_96":180,"_159":181,"_161":-5,"_162":144,"_163":182,"_165":166,"_107":183,"_101":168,"_169":170,"_171":184,"_173":22,"_174":-5,"_175":402,"_99":403},[],[180],{"_105":188,"_96":189,"_159":190,"_161":-5,"_162":144,"_163":191,"_165":166,"_107":192,"_101":193,"_169":170,"_171":194,"_173":22,"_174":-5,"_175":405,"_99":406},[],[189],{"_105":198,"_96":199,"_159":200,"_161":-5,"_162":144,"_163":201,"_165":166,"_107":202,"_101":193,"_169":170,"_171":203,"_173":22,"_174":-5,"_175":408,"_99":409},[],[199],{"_105":207,"_96":100,"_159":208,"_161":-5,"_162":144,"_163":209,"_165":166,"_107":210,"_101":211,"_169":170,"_171":212,"_173":22,"_174":213,"_175":411,"_99":412},[],[100],{"_105":217,"_96":218,"_159":219,"_161":-5,"_162":144,"_163":220,"_165":166,"_107":221,"_101":193,"_169":170,"_171":222,"_173":22,"_174":-5,"_175":414,"_99":415},[],[218],{"_105":226,"_96":227,"_159":228,"_161":-5,"_162":144,"_163":229,"_165":166,"_107":230,"_101":193,"_169":170,"_171":203,"_173":22,"_174":-5,"_175":417,"_99":418},[],[227],{"_105":234,"_96":235,"_159":236,"_161":-5,"_162":144,"_163":237,"_165":166,"_107":238,"_101":239,"_169":170,"_171":240,"_173":22,"_174":-5,"_175":420,"_99":421},[],[235],{"_105":244,"_96":245,"_159":246,"_161":-5,"_162":144,"_163":247,"_165":166,"_107":248,"_101":249,"_169":170,"_171":250,"_173":22,"_174":251,"_175":423,"_99":424},[],[245],{"_105":255,"_96":256,"_159":257,"_161":-5,"_162":144,"_163":258,"_165":166,"_107":259,"_101":168,"_169":170,"_171":260,"_173":22,"_174":-5,"_175":426,"_99":427},[],[256],{"_105":264,"_96":265,"_159":266,"_161":-5,"_162":144,"_163":267,"_165":166,"_107":268,"_101":269,"_169":170,"_171":270,"_173":22,"_174":-5,"_175":429,"_99":430},[],[265],{"_105":274,"_96":275,"_159":276,"_161":-5,"_162":144,"_163":277,"_165":166,"_107":278,"_101":279,"_169":170,"_171":280,"_173":22,"_174":-5,"_175":432,"_99":433},[],[275],{"_105":284,"_96":285,"_159":286,"_161":-5,"_162":144,"_163":287,"_165":166,"_107":288,"_101":168,"_169":170,"_171":289,"_173":22,"_174":-5,"_175":435,"_99":436},[],[285],{"_105":293,"_96":294,"_159":295,"_161":-5,"_162":144,"_163":296,"_165":166,"_107":297,"_101":193,"_169":170,"_171":298,"_173":22,"_174":-5,"_175":438,"_99":439},[],[294],{"_105":302,"_96":303,"_159":304,"_161":-5,"_162":144,"_163":305,"_165":166,"_107":306,"_101":249,"_169":170,"_171":307,"_173":22,"_174":-5,"_175":441,"_99":442},[],[303],{"_105":311,"_96":312,"_159":313,"_161":-5,"_162":144,"_163":314,"_165":166,"_107":315,"_101":32,"_169":316,"_171":73,"_173":22,"_174":-5,"_175":444,"_99":445},[],[312],{"_105":320,"_96":321,"_159":322,"_161":-5,"_162":144,"_163":102,"_165":22,"_107":323,"_101":258,"_169":170,"_171":324,"_173":22,"_174":-5,"_175":447,"_99":448},[],[321],{"_105":328,"_96":329,"_159":330,"_161":-5,"_162":144,"_163":331,"_165":166,"_107":332,"_101":279,"_169":170,"_171":333,"_173":22,"_174":-5,"_175":450,"_99":451},[],[329],{"_105":337,"_96":338,"_159":339,"_161":-5,"_162":144,"_163":269,"_165":166,"_107":340,"_101":279,"_169":170,"_171":341,"_173":22,"_174":-5,"_175":453,"_99":454},[],[338],{"_105":345,"_96":346,"_159":347,"_161":-5,"_162":144,"_163":348,"_165":166,"_107":349,"_101":350,"_169":170,"_171":351,"_173":22,"_174":-5,"_175":456,"_99":457},[],[346],{"_105":355,"_96":356,"_159":357,"_161":-5,"_162":144,"_163":358,"_165":166,"_107":359,"_101":168,"_169":170,"_171":360,"_173":22,"_174":-5,"_175":459,"_99":460},[],[356],[398,401,404,407,410,413,416,419,422,425,428,431,434,437,440,443,446,449,452,455,458],{"_105":207,"_96":100,"_159":208,"_161":-5,"_162":144,"_163":209,"_165":166,"_107":382,"_101":211,"_169":170,"_171":212,"_173":22,"_174":213,"_99":463},[100],[],"upstream","pull","ahead","behind","templates",[],"actionData","errors"]

ziggyware

agentHub