[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"summary-10eae276fc8f2aed-adk-build-production-ai-agents-at-scale-summary":3,"summaries-facets-categories":115,"summary-related-10eae276fc8f2aed-adk-build-production-ai-agents-at-scale-summary":3685},{"id":4,"title":5,"ai":6,"body":13,"categories":78,"created_at":79,"date_modified":79,"description":72,"extension":80,"faq":79,"featured":81,"kicker_label":79,"meta":82,"navigation":97,"path":98,"published_at":79,"question":79,"scraped_at":99,"seo":100,"sitemap":101,"source_id":102,"source_name":103,"source_type":104,"source_url":105,"stem":106,"tags":107,"thumbnail_url":79,"tldr":112,"tweet":79,"unknown_tags":113,"__hash__":114},"summaries\u002Fsummaries\u002F10eae276fc8f2aed-adk-build-production-ai-agents-at-scale-summary.md","ADK: Build Production AI Agents at Scale",{"provider":7,"model":8,"input_tokens":9,"output_tokens":10,"processing_time_ms":11,"cost_usd":12},"openrouter","x-ai\u002Fgrok-4.1-fast",8114,1769,7685,0.00200045,{"type":14,"value":15,"toc":71},"minimark",[16,21,57,61,64,68],[17,18,20],"h2",{"id":19},"define-agents-with-minimal-code-for-immediate-use","Define Agents with Minimal Code for Immediate Use",[22,23,24,25,29,30,33,34,37,38,41,42,45,46,45,49,52,53,56],"p",{},"Create functional LLM agents using a single class instantiation across languages, specifying name, model (e.g., gemini-flash-latest), instruction, and tools like google_search. In Python: ",[26,27,28],"code",{},"from google.adk import Agent; agent = Agent(name=\"researcher\", model=\"gemini-flash-latest\", instruction=\"You help users research topics thoroughly.\", tools=[google_search])",". TypeScript uses ",[26,31,32],{},"LlmAgent"," constructor similarly; Go uses ",[26,35,36],{},"agent.New"," with options; Java uses ",[26,39,40],{},"LlmAgent.builder()",". Install via ",[26,43,44],{},"pip install google-adk",", ",[26,47,48],{},"npm install @google\u002Fadk",[26,50,51],{},"go get google.golang.org\u002Fadk",", or Maven ",[26,54,55],{},"com.google.adk:google-adk",". This approach scales from simple tool-calling agents to multi-agent systems, workflow agents (sequential, loop, parallel), and custom agents without initial complexity.",[17,58,60],{"id":59},"manage-context-like-source-code-for-efficiency","Manage Context Like Source Code for Efficiency",[22,62,63],{},"ADK structures context from sessions, memory, tool outputs, and artifacts, filtering irrelevant events, summarizing old turns, lazy-loading artifacts, and tracking tokens to avoid overflow and keep agents fast. Customize via caching, compression, and compaction. Sessions support rewind and migration; state and memory persist across runs. Use callbacks for event interception, artifacts for generated content, and events for observability. This prevents the common pitfall of concatenating strings until failure, ensuring reliability in long-running tasks.",[17,65,67],{"id":66},"evaluate-deploy-and-integrate-for-production","Evaluate, Deploy, and Integrate for Production",[22,69,70],{},"Test agents with visual debugging, user\u002Fenvironment simulation, custom metrics, and optimization loops. Deploy via containerization anywhere or one-command to Google Cloud's Agent Engine (inherits auth, tracing, security), Cloud Run, or GKE without code changes. Run via web UI, CLI, API server, or resume interrupted sessions. Supports models like Gemini, Gemma, Claude, Vertex AI, Ollama, vLLM, LiteLLM; tools including function, MCP, OpenAPI; integrations for apps, plugins, grounding (Google\u002FVertex Search), and A2A protocol for agent-to-agent communication. Build multi-agent teams, graph-based workflows (routes, data handling, human input), and streaming with Gemini Live API Toolkit handling audio\u002Fimages\u002Fvideo.",{"title":72,"searchDepth":73,"depth":73,"links":74},"",2,[75,76,77],{"id":19,"depth":73,"text":20},{"id":59,"depth":73,"text":60},{"id":66,"depth":73,"text":67},[],null,"md",false,{"content_references":83,"triage":92},[84,89],{"type":85,"title":86,"url":87,"context":88},"other","ADK Go 1.0","https:\u002F\u002Fdevelopers.googleblog.com\u002Fadk-go-10-arrives\u002F","mentioned",{"type":85,"title":90,"url":91,"context":88},"ADK Java 1.0","https:\u002F\u002Fdevelopers.googleblog.com\u002Fannouncing-adk-for-java-100-building-the-future-of-ai-agents-in-java\u002F",{"relevance":93,"novelty":94,"quality":94,"actionability":93,"composite":95,"reasoning":96},5,4,4.55,"Category: AI & LLMs. The article provides a comprehensive overview of the ADK framework for building AI agents, addressing practical applications and specific pain points for developers looking to implement AI features in production. It includes actionable code examples and deployment strategies, making it immediately useful for the target audience.",true,"\u002Fsummaries\u002F10eae276fc8f2aed-adk-build-production-ai-agents-at-scale-summary","2026-04-16 03:06:19",{"title":5,"description":72},{"loc":98},"10eae276fc8f2aed","__oneoff__","article","https:\u002F\u002Fgoogle.github.io\u002Fadk-docs\u002F","summaries\u002F10eae276fc8f2aed-adk-build-production-ai-agents-at-scale-summary",[108,109,110,111],"agents","llm","ai-tools","open-source","Google's open-source ADK framework enables building reliable AI agents in Python, TypeScript, Go, Java with structured context management, multi-model support, evaluation tools, and seamless Google Cloud deployment.",[],"31cqQWE3k-a4cF0poOakxirWEb0b_-p4BW1f1nqFOG4",[116,119,122,125,128,131,133,135,137,139,141,143,146,148,150,152,154,156,158,160,162,164,167,170,172,174,177,179,181,184,186,188,190,192,194,196,198,200,202,204,206,208,210,212,214,216,218,220,222,224,226,228,230,232,234,236,238,240,242,244,246,248,250,252,254,256,258,260,262,264,266,268,270,272,274,276,278,280,282,284,286,288,290,292,294,296,298,300,302,304,306,308,310,312,314,316,318,320,322,324,326,328,330,332,334,336,338,340,342,344,346,348,350,352,354,356,358,360,362,364,366,368,370,372,374,376,378,380,382,384,386,388,390,392,394,396,398,400,402,404,406,408,410,412,414,416,418,420,422,424,426,428,430,432,434,436,439,441,443,445,447,449,451,453,455,457,459,461,463,465,467,469,471,473,475,477,479,481,483,485,487,489,491,493,495,497,499,501,503,505,507,509,511,513,515,517,519,521,523,525,527,529,531,533,535,537,539,541,543,545,547,549,551,553,555,557,559,561,563,565,567,569,571,573,575,577,579,581,583,585,587,589,591,593,595,597,599,601,603,605,607,609,611,613,615,617,619,621,623,625,627,629,631,633,635,637,639,641,643,645,647,649,651,653,655,657,659,661,663,665,667,669,671,673,675,677,679,681,683,685,687,689,691,693,695,697,699,701,703,705,707,709,711,713,715,717,719,721,723,725,727,729,731,733,735,737,739,741,743,745,747,749,751,753,755,757,759,761,763,765,767,769,771,773,775,777,779,781,783,785,787,789,791,793,795,797,799,801,803,805,807,809,811,813,815,817,819,821,823,825,827,829,831,833,835,837,839,841,843,845,847,849,851,853,855,857,859,861,863,865,867,869,871,873,875,877,879,881,883,885,887,889,891,893,895,897,899,901,903,905,907,909,911,913,915,917,919,921,923,925,927,929,931,933,935,937,939,941,943,945,947,949,951,953,955,957,959,961,963,965,967,969,971,973,975,977,979,981,983,985,987,989,991,993,995,997,999,1001,1003,1005,1007,1009,1011,1013,1015,1017,1019,1021,1023,1025,1027,1029,1031,1033,1035,1037,1039,1041,1043,1045,1047,1049,1051,1053,1055,1057,1059,1061,1063,1065,1067,1069,1071,1073,1075,1077,1079,1081,1083,1085,1087,1089,1091,1093,1095,1097,1099,1101,1103,1105,1107,1109,1111,1113,1115,1117,1119,1121,1123,1125,1127,1129,1131,1133,1135,1137,1139,1141,1143,1145,1147,1149,1151,1153,1155,1157,1159,1161,1163,1165,1167,1169,1171,1173,1175,1177,1179,1181,1183,1185,1187,1189,1191,1193,1195,1197,1199,1201,1203,1205,1207,1209,1211,1213,1215,1217,1219,1221,1223,1225,1227,1229,1231,1233,1235,1237,1239,1241,1243,1245,1247,1249,1251,1253,1255,1257,1259,1261,1263,1265,1267,1269,1271,1273,1275,1277,1279,1281,1283,1285,1287,1289,1291,1293,1295,1297,1299,1301,1303,1305,1307,1309,1311,1313,1315,1317,1319,1321,1323,1325,1327,1329,1331,1333,1335,1337,1339,1341,1343,1345,1347,1349,1351,1353,1355,1357,1359,1361,1363,1365,1367,1369,1371,1373,1375,1377,1379,1381,1383,1385,1387,1389,1391,1393,1395,1397,1399,1401,1403,1405,1407,1409,1411,1413,1415,1417,1419,1421,1423,1425,1427,1429,1431,1433,1435,1437,1439,1441,1443,1445,1447,1449,1451,1453,1455,1457,1459,1461,1463,1465,1467,1469,1471,1473,1475,1477,1479,1481,1483,1485,1487,1489,1491,1493,1495,1497,1499,1501,1503,1505,1507,1509,1511,1513,1515,1517,1519,1521,1523,1525,1527,1529,1531,1533,1535,1537,1539,1541,1543,1545,1547,1549,1551,1553,1555,1557,1559,1561,1563,1565,1567,1569,1571,1573,1575,1577,1579,1581,1583,1585,1587,1589,1591,1593,1595,1597,1599,1601,1603,1605,1607,1609,1611,1613,1615,1617,1619,1621,1623,1625,1627,1629,1631,1633,1635,1637,1639,1641,1643,1645,1647,1649,1651,1653,1655,1657,1659,1661,1663,1665,1667,1669,1671,1673,1675,1677,1679,1681,1683,1685,1687,1689,1691,1693,1695,1697,1699,1701,1703,1705,1707,1709,1711,1713,1715,1717,1719,1721,1723,1725,1727,1729,1731,1733,1735,1737,1739,1741,1743,1745,1747,1749,1751,1753,1755,1757,1759,1761,1763,1765,1767,1769,1771,1773,1775,1777,1779,1781,1783,1785,1787,1789,1791,1793,1795,1797,1799,1801,1803,1805,1807,1809,1811,1813,1815,1817,1819,1821,1823,1825,1827,1829,1831,1833,1835,1837,1839,1841,1843,1845,1847,1849,1851,1853,1855,1857,1859,1861,1863,1865,1867,1869,1871,1873,1875,1877,1879,1881,1883,1885,1887,1889,1891,1893,1895,1897,1899,1901,1903,1905,1907,1909,1911,1913,1915,1917,1919,1921,1923,1925,1927,1929,1931,1933,1935,1937,1939,1941,1943,1945,1947,1949,1951,1953,1955,1957,1959,1961,1963,1965,1967,1969,1971,1973,1975,1977,1979,1981,1983,1985,1987,1989,1991,1993,1995,1997,1999,2001,2003,2005,2007,2009,2011,2013,2015,2017,2019,2021,2023,2025,2027,2029,2031,2033,2035,2037,2039,2041,2043,2045,2047,2049,2051,2053,2055,2057,2059,2061,2063,2065,2067,2069,2071,2073,2075,2077,2079,2081,2083,2085,2087,2089,2091,2093,2095,2097,2099,2101,2103,2105,2107,2109,2111,2113,2115,2117,2119,2121,2123,2125,2127,2129,2131,2133,2135,2137,2139,2141,2143,2145,2147,2149,2151,2153,2155,2157,2159,2161,2163,2165,2167,2169,2171,2173,2175,2177,2179,2181,2183,2185,2187,2189,2191,2193,2195,2197,2199,2201,2203,2205,2207,2209,2211,2213,2215,2217,2219,2221,2223,2225,2227,2229,2231,2233,2235,2237,2239,2241,2243,2245,2247,2249,2251,2253,2255,2257,2259,2261,2263,2265,2267,2269,2271,2273,2275,2277,2279,2281,2283,2285,2287,2289,2291,2293,2295,2297,2299,2301,2303,2305,2307,2309,2311,2313,2315,2317,2319,2321,2323,2325,2327,2329,2331,2333,2335,2337,2339,2341,2343,2345,2347,2349,2351,2353,2355,2357,2359,2361,2363,2365,2367,2369,2371,2373,2375,2377,2379,2381,2383,2385,2387,2389,2391,2393,2395,2397,2399,2401,2403,2405,2407,2409,2411,2413,2415,2417,2419,2421,2423,2425,2427,2429,2431,2433,2435,2437,2439,2441,2443,2445,2447,2449,2451,2453,2455,2457,2459,2461,2463,2465,2467,2469,2471,2473,2475,2477,2479,2481,2483,2485,2487,2489,2491,2493,2495,2497,2499,2501,2503,2505,2507,2509,2511,2513,2515,2517,2519,2521,2523,2525,2527,2529,2531,2533,2535,2537,2539,2541,2543,2545,2547,2549,2551,2553,2555,2557,2559,2561,2563,2565,2567,2569,2571,2573,2575,2577,2579,2581,2583,2585,2587,2589,2591,2593,2595,2597,2599,2601,2603,2605,2607,2609,2611,2613,2615,2617,2619,2621,2623,2625,2627,2629,2631,2633,2635,2637,2639,2641,2643,2645,2647,2649,2651,2653,2655,2657,2659,2661,2663,2665,2667,2669,2671,2673,2675,2677,2679,2681,2683,2685,2687,2689,2691,2693,2695,2697,2699,2701,2703,2705,2707,2709,2711,2713,2715,2717,2719,2721,2723,2725,2727,2729,2731,2733,2735,2737,2739,2741,2743,2745,2747,2749,2751,2753,2755,2757,2759,2761,2763,2765,2767,2769,2771,2773,2775,2777,2779,2781,2783,2785,2787,2789,2791,2793,2795,2797,2799,2801,2803,2805,2807,2809,2811,2813,2815,2817,2819,2821,2823,2825,2827,2829,2831,2833,2835,2837,2839,2841,2843,2845,2847,2849,2851,2853,2855,2857,2859,2861,2863,2865,2867,2869,2871,2873,2875,2877,2879,2881,2883,2885,2887,2889,2891,2893,2895,2897,2899,2901,2903,2905,2907,2909,2911,2913,2915,2917,2919,2921,2923,2925,2927,2929,2931,2933,2935,2937,2939,2941,2943,2945,2947,2949,2951,2953,2955,2957,2959,2961,2963,2965,2967,2969,2971,2973,2975,2977,2979,2981,2983,2985,2987,2989,2991,2993,2995,2997,2999,3001,3003,3005,3007,3009,3011,3013,3015,3017,3019,3021,3023,3025,3027,3029,3031,3033,3035,3037,3039,3041,3043,3045,3047,3049,3051,3053,3055,3057,3059,3061,3063,3065,3067,3069,3071,3073,3075,3077,3079,3081,3083,3085,3087,3089,3091,3093,3095,3097,3099,3101,3103,3105,3107,3109,3111,3113,3115,3117,3119,3121,3123,3125,3127,3129,3131,3133,3135,3137,3139,3141,3143,3145,3147,3149,3151,3153,3155,3157,3159,3161,3163,3165,3167,3169,3171,3173,3175,3177,3179,3181,3183,3185,3187,3189,3191,3193,3195,3197,3199,3201,3203,3205,3207,3209,3211,3213,3215,3217,3219,3221,3223,3225,3227,3229,3231,3233,3235,3237,3239,3241,3243,3245,3247,3249,3251,3253,3255,3257,3259,3261,3263,3265,3267,3269,3271,3273,3275,3277,3279,3281,3283,3285,3287,3289,3291,3293,3295,3297,3299,3301,3303,3305,3307,3309,3311,3313,3315,3317,3319,3321,3323,3325,3327,3329,3331,3333,3335,3337,3339,3341,3343,3345,3347,3349,3351,3353,3355,3357,3359,3361,3363,3365,3367,3369,3371,3373,3375,3377,3379,3381,3383,3385,3387,3389,3391,3393,3395,3397,3399,3401,3403,3405,3407,3409,3411,3413,3415,3417,3419,3421,3423,3425,3427,3429,3431,3433,3435,3437,3439,3441,3443,3445,3447,3449,3451,3453,3455,3457,3459,3461,3463,3465,3467,3469,3471,3473,3475,3477,3479,3481,3483,3485,3487,3489,3491,3493,3495,3497,3499,3501,3503,3505,3507,3509,3511,3513,3515,3517,3519,3521,3523,3525,3527,3529,3531,3533,3535,3537,3539,3541,3543,3545,3547,3549,3551,3553,3555,3557,3559,3561,3563,3565,3567,3569,3571,3573,3575,3577,3579,3581,3583,3585,3587,3589,3591,3593,3595,3597,3599,3601,3603,3605,3607,3609,3611,3613,3615,3617,3619,3621,3623,3625,3627,3629,3631,3633,3635,3637,3639,3641,3643,3645,3647,3649,3651,3653,3655,3657,3659,3661,3663,3665,3667,3669,3671,3673,3675,3677,3679,3681,3683],{"categories":117},[118],"Developer Productivity",{"categories":120},[121],"Business & SaaS",{"categories":123},[124],"AI & LLMs",{"categories":126},[127],"AI Automation",{"categories":129},[130],"Product Strategy",{"categories":132},[124],{"categories":134},[118],{"categories":136},[121],{"categories":138},[],{"categories":140},[124],{"categories":142},[],{"categories":144},[145],"AI News & Trends",{"categories":147},[127],{"categories":149},[145],{"categories":151},[127],{"categories":153},[127],{"categories":155},[124],{"categories":157},[124],{"categories":159},[145],{"categories":161},[124],{"categories":163},[],{"categories":165},[166],"Design & Frontend",{"categories":168},[169],"Data Science & Visualization",{"categories":171},[145],{"categories":173},[],{"categories":175},[176],"Software Engineering",{"categories":178},[124],{"categories":180},[127],{"categories":182},[183],"Marketing & Growth",{"categories":185},[124],{"categories":187},[127],{"categories":189},[],{"categories":191},[],{"categories":193},[166],{"categories":195},[127],{"categories":197},[118],{"categories":199},[166],{"categories":201},[124],{"categories":203},[127],{"categories":205},[145],{"categories":207},[],{"categories":209},[],{"categories":211},[127],{"categories":213},[176],{"categories":215},[],{"categories":217},[121],{"categories":219},[],{"categories":221},[],{"categories":223},[127],{"categories":225},[127],{"categories":227},[124],{"categories":229},[],{"categories":231},[176],{"categories":233},[],{"categories":235},[],{"categories":237},[],{"categories":239},[124],{"categories":241},[183],{"categories":243},[166],{"categories":245},[166],{"categories":247},[124],{"categories":249},[127],{"categories":251},[124],{"categories":253},[124],{"categories":255},[127],{"categories":257},[127],{"categories":259},[169],{"categories":261},[145],{"categories":263},[127],{"categories":265},[183],{"categories":267},[127],{"categories":269},[130],{"categories":271},[],{"categories":273},[127],{"categories":275},[],{"categories":277},[127],{"categories":279},[176],{"categories":281},[166],{"categories":283},[124],{"categories":285},[],{"categories":287},[],{"categories":289},[127],{"categories":291},[],{"categories":293},[124],{"categories":295},[],{"categories":297},[118],{"categories":299},[176],{"categories":301},[121],{"categories":303},[145],{"categories":305},[124],{"categories":307},[],{"categories":309},[124],{"categories":311},[],{"categories":313},[176],{"categories":315},[169],{"categories":317},[],{"categories":319},[124],{"categories":321},[166],{"categories":323},[],{"categories":325},[166],{"categories":327},[127],{"categories":329},[],{"categories":331},[127],{"categories":333},[145],{"categories":335},[124],{"categories":337},[],{"categories":339},[127],{"categories":341},[124],{"categories":343},[130],{"categories":345},[],{"categories":347},[124],{"categories":349},[127],{"categories":351},[127],{"categories":353},[],{"categories":355},[169],{"categories":357},[124],{"categories":359},[],{"categories":361},[118],{"categories":363},[121],{"categories":365},[124],{"categories":367},[127],{"categories":369},[176],{"categories":371},[124],{"categories":373},[],{"categories":375},[],{"categories":377},[124],{"categories":379},[],{"categories":381},[166],{"categories":383},[],{"categories":385},[124],{"categories":387},[],{"categories":389},[127],{"categories":391},[124],{"categories":393},[166],{"categories":395},[],{"categories":397},[124],{"categories":399},[124],{"categories":401},[121],{"categories":403},[127],{"categories":405},[124],{"categories":407},[166],{"categories":409},[127],{"categories":411},[],{"categories":413},[],{"categories":415},[145],{"categories":417},[],{"categories":419},[124],{"categories":421},[121,183],{"categories":423},[],{"categories":425},[124],{"categories":427},[],{"categories":429},[],{"categories":431},[124],{"categories":433},[],{"categories":435},[124],{"categories":437},[438],"DevOps & Cloud",{"categories":440},[],{"categories":442},[145],{"categories":444},[166],{"categories":446},[],{"categories":448},[145],{"categories":450},[145],{"categories":452},[124],{"categories":454},[183],{"categories":456},[],{"categories":458},[121],{"categories":460},[],{"categories":462},[124,438],{"categories":464},[124],{"categories":466},[124],{"categories":468},[127],{"categories":470},[124,176],{"categories":472},[169],{"categories":474},[124],{"categories":476},[183],{"categories":478},[127],{"categories":480},[127],{"categories":482},[],{"categories":484},[127],{"categories":486},[124,121],{"categories":488},[],{"categories":490},[166],{"categories":492},[166],{"categories":494},[],{"categories":496},[],{"categories":498},[145],{"categories":500},[],{"categories":502},[118],{"categories":504},[176],{"categories":506},[124],{"categories":508},[166],{"categories":510},[127],{"categories":512},[176],{"categories":514},[145],{"categories":516},[166],{"categories":518},[],{"categories":520},[124],{"categories":522},[124],{"categories":524},[124],{"categories":526},[145],{"categories":528},[118],{"categories":530},[124],{"categories":532},[127],{"categories":534},[438],{"categories":536},[166],{"categories":538},[127],{"categories":540},[],{"categories":542},[],{"categories":544},[166],{"categories":546},[145],{"categories":548},[169],{"categories":550},[],{"categories":552},[124],{"categories":554},[124],{"categories":556},[121],{"categories":558},[124],{"categories":560},[124],{"categories":562},[145],{"categories":564},[],{"categories":566},[127],{"categories":568},[176],{"categories":570},[],{"categories":572},[124],{"categories":574},[124],{"categories":576},[127],{"categories":578},[],{"categories":580},[],{"categories":582},[124],{"categories":584},[],{"categories":586},[121],{"categories":588},[127],{"categories":590},[],{"categories":592},[118],{"categories":594},[124],{"categories":596},[121],{"categories":598},[145],{"categories":600},[],{"categories":602},[],{"categories":604},[],{"categories":606},[145],{"categories":608},[145],{"categories":610},[],{"categories":612},[],{"categories":614},[121],{"categories":616},[],{"categories":618},[],{"categories":620},[118],{"categories":622},[],{"categories":624},[183],{"categories":626},[127],{"categories":628},[121],{"categories":630},[127],{"categories":632},[],{"categories":634},[130],{"categories":636},[166],{"categories":638},[176],{"categories":640},[124],{"categories":642},[127],{"categories":644},[121],{"categories":646},[124],{"categories":648},[],{"categories":650},[],{"categories":652},[176],{"categories":654},[169],{"categories":656},[130],{"categories":658},[127],{"categories":660},[124],{"categories":662},[],{"categories":664},[438],{"categories":666},[],{"categories":668},[127],{"categories":670},[],{"categories":672},[],{"categories":674},[124],{"categories":676},[166],{"categories":678},[183],{"categories":680},[127],{"categories":682},[],{"categories":684},[118],{"categories":686},[],{"categories":688},[145],{"categories":690},[124,438],{"categories":692},[145],{"categories":694},[124],{"categories":696},[121],{"categories":698},[124],{"categories":700},[],{"categories":702},[121],{"categories":704},[],{"categories":706},[176],{"categories":708},[166],{"categories":710},[145],{"categories":712},[169],{"categories":714},[118],{"categories":716},[124],{"categories":718},[176],{"categories":720},[],{"categories":722},[],{"categories":724},[130],{"categories":726},[],{"categories":728},[124],{"categories":730},[],{"categories":732},[166],{"categories":734},[166],{"categories":736},[166],{"categories":738},[],{"categories":740},[],{"categories":742},[145],{"categories":744},[127],{"categories":746},[124],{"categories":748},[124],{"categories":750},[124],{"categories":752},[121],{"categories":754},[124],{"categories":756},[],{"categories":758},[176],{"categories":760},[176],{"categories":762},[121],{"categories":764},[],{"categories":766},[124],{"categories":768},[124],{"categories":770},[121],{"categories":772},[145],{"categories":774},[183],{"categories":776},[127],{"categories":778},[],{"categories":780},[166],{"categories":782},[],{"categories":784},[124],{"categories":786},[],{"categories":788},[121],{"categories":790},[127],{"categories":792},[],{"categories":794},[438],{"categories":796},[169],{"categories":798},[176],{"categories":800},[183],{"categories":802},[176],{"categories":804},[127],{"categories":806},[],{"categories":808},[],{"categories":810},[127],{"categories":812},[118],{"categories":814},[127],{"categories":816},[130],{"categories":818},[121],{"categories":820},[],{"categories":822},[124],{"categories":824},[130],{"categories":826},[124],{"categories":828},[124],{"categories":830},[183],{"categories":832},[166],{"categories":834},[127],{"categories":836},[],{"categories":838},[],{"categories":840},[438],{"categories":842},[176],{"categories":844},[],{"categories":846},[127],{"categories":848},[124],{"categories":850},[166,124],{"categories":852},[118],{"categories":854},[],{"categories":856},[124],{"categories":858},[118],{"categories":860},[166],{"categories":862},[127],{"categories":864},[176],{"categories":866},[],{"categories":868},[124],{"categories":870},[],{"categories":872},[118],{"categories":874},[],{"categories":876},[127],{"categories":878},[130],{"categories":880},[124],{"categories":882},[124],{"categories":884},[166],{"categories":886},[127],{"categories":888},[438],{"categories":890},[166],{"categories":892},[127],{"categories":894},[124],{"categories":896},[124],{"categories":898},[124],{"categories":900},[145],{"categories":902},[],{"categories":904},[130],{"categories":906},[127],{"categories":908},[166],{"categories":910},[127],{"categories":912},[176],{"categories":914},[166],{"categories":916},[127],{"categories":918},[145],{"categories":920},[],{"categories":922},[124],{"categories":924},[166],{"categories":926},[124],{"categories":928},[118],{"categories":930},[145],{"categories":932},[124],{"categories":934},[183],{"categories":936},[124],{"categories":938},[124],{"categories":940},[127],{"categories":942},[127],{"categories":944},[124],{"categories":946},[127],{"categories":948},[166],{"categories":950},[124],{"categories":952},[],{"categories":954},[],{"categories":956},[176],{"categories":958},[],{"categories":960},[118],{"categories":962},[438],{"categories":964},[],{"categories":966},[118],{"categories":968},[121],{"categories":970},[183],{"categories":972},[],{"categories":974},[121],{"categories":976},[],{"categories":978},[],{"categories":980},[],{"categories":982},[],{"categories":984},[],{"categories":986},[124],{"categories":988},[127],{"categories":990},[438],{"categories":992},[118],{"categories":994},[124],{"categories":996},[176],{"categories":998},[130],{"categories":1000},[124],{"categories":1002},[183],{"categories":1004},[124],{"categories":1006},[124],{"categories":1008},[124],{"categories":1010},[124,118],{"categories":1012},[176],{"categories":1014},[176],{"categories":1016},[166],{"categories":1018},[124],{"categories":1020},[],{"categories":1022},[],{"categories":1024},[],{"categories":1026},[176],{"categories":1028},[169],{"categories":1030},[145],{"categories":1032},[166],{"categories":1034},[],{"categories":1036},[124],{"categories":1038},[124],{"categories":1040},[],{"categories":1042},[],{"categories":1044},[127],{"categories":1046},[124],{"categories":1048},[121],{"categories":1050},[],{"categories":1052},[118],{"categories":1054},[124],{"categories":1056},[118],{"categories":1058},[124],{"categories":1060},[176],{"categories":1062},[183],{"categories":1064},[124,166],{"categories":1066},[145],{"categories":1068},[166],{"categories":1070},[],{"categories":1072},[438],{"categories":1074},[166],{"categories":1076},[127],{"categories":1078},[],{"categories":1080},[],{"categories":1082},[],{"categories":1084},[],{"categories":1086},[176],{"categories":1088},[127],{"categories":1090},[127],{"categories":1092},[124],{"categories":1094},[124],{"categories":1096},[],{"categories":1098},[166],{"categories":1100},[],{"categories":1102},[],{"categories":1104},[127],{"categories":1106},[],{"categories":1108},[],{"categories":1110},[183],{"categories":1112},[183],{"categories":1114},[127],{"categories":1116},[],{"categories":1118},[124],{"categories":1120},[124],{"categories":1122},[176],{"categories":1124},[166],{"categories":1126},[166],{"categories":1128},[127],{"categories":1130},[118],{"categories":1132},[124],{"categories":1134},[166],{"categories":1136},[166],{"categories":1138},[127],{"categories":1140},[127],{"categories":1142},[124],{"categories":1144},[],{"categories":1146},[],{"categories":1148},[124],{"categories":1150},[127],{"categories":1152},[145],{"categories":1154},[176],{"categories":1156},[118],{"categories":1158},[124],{"categories":1160},[],{"categories":1162},[127],{"categories":1164},[127],{"categories":1166},[],{"categories":1168},[118],{"categories":1170},[124],{"categories":1172},[118],{"categories":1174},[118],{"categories":1176},[],{"categories":1178},[],{"categories":1180},[127],{"categories":1182},[127],{"categories":1184},[124],{"categories":1186},[124],{"categories":1188},[145],{"categories":1190},[169],{"categories":1192},[130],{"categories":1194},[145],{"categories":1196},[166],{"categories":1198},[],{"categories":1200},[145],{"categories":1202},[],{"categories":1204},[],{"categories":1206},[],{"categories":1208},[],{"categories":1210},[176],{"categories":1212},[169],{"categories":1214},[],{"categories":1216},[124],{"categories":1218},[124],{"categories":1220},[169],{"categories":1222},[176],{"categories":1224},[],{"categories":1226},[],{"categories":1228},[127],{"categories":1230},[145],{"categories":1232},[145],{"categories":1234},[127],{"categories":1236},[118],{"categories":1238},[124,438],{"categories":1240},[],{"categories":1242},[166],{"categories":1244},[118],{"categories":1246},[127],{"categories":1248},[166],{"categories":1250},[],{"categories":1252},[127],{"categories":1254},[127],{"categories":1256},[124],{"categories":1258},[183],{"categories":1260},[176],{"categories":1262},[166],{"categories":1264},[],{"categories":1266},[127],{"categories":1268},[124],{"categories":1270},[127],{"categories":1272},[127],{"categories":1274},[127],{"categories":1276},[183],{"categories":1278},[127],{"categories":1280},[124],{"categories":1282},[],{"categories":1284},[183],{"categories":1286},[145],{"categories":1288},[127],{"categories":1290},[],{"categories":1292},[],{"categories":1294},[124],{"categories":1296},[127],{"categories":1298},[145],{"categories":1300},[127],{"categories":1302},[],{"categories":1304},[],{"categories":1306},[],{"categories":1308},[127],{"categories":1310},[],{"categories":1312},[],{"categories":1314},[169],{"categories":1316},[124],{"categories":1318},[169],{"categories":1320},[145],{"categories":1322},[124],{"categories":1324},[124],{"categories":1326},[127],{"categories":1328},[124],{"categories":1330},[],{"categories":1332},[],{"categories":1334},[438],{"categories":1336},[],{"categories":1338},[],{"categories":1340},[118],{"categories":1342},[],{"categories":1344},[],{"categories":1346},[],{"categories":1348},[],{"categories":1350},[176],{"categories":1352},[145],{"categories":1354},[183],{"categories":1356},[121],{"categories":1358},[124],{"categories":1360},[124],{"categories":1362},[121],{"categories":1364},[],{"categories":1366},[166],{"categories":1368},[127],{"categories":1370},[121],{"categories":1372},[124],{"categories":1374},[124],{"categories":1376},[118],{"categories":1378},[],{"categories":1380},[118],{"categories":1382},[124],{"categories":1384},[183],{"categories":1386},[127],{"categories":1388},[145],{"categories":1390},[121],{"categories":1392},[124],{"categories":1394},[127],{"categories":1396},[],{"categories":1398},[124],{"categories":1400},[118],{"categories":1402},[124],{"categories":1404},[],{"categories":1406},[145],{"categories":1408},[124],{"categories":1410},[],{"categories":1412},[121],{"categories":1414},[124],{"categories":1416},[],{"categories":1418},[],{"categories":1420},[],{"categories":1422},[124],{"categories":1424},[],{"categories":1426},[438],{"categories":1428},[124],{"categories":1430},[],{"categories":1432},[124],{"categories":1434},[124],{"categories":1436},[124],{"categories":1438},[124,438],{"categories":1440},[124],{"categories":1442},[124],{"categories":1444},[166],{"categories":1446},[127],{"categories":1448},[],{"categories":1450},[127],{"categories":1452},[124],{"categories":1454},[124],{"categories":1456},[124],{"categories":1458},[118],{"categories":1460},[118],{"categories":1462},[176],{"categories":1464},[166],{"categories":1466},[127],{"categories":1468},[],{"categories":1470},[124],{"categories":1472},[145],{"categories":1474},[124],{"categories":1476},[121],{"categories":1478},[],{"categories":1480},[438],{"categories":1482},[166],{"categories":1484},[166],{"categories":1486},[127],{"categories":1488},[145],{"categories":1490},[127],{"categories":1492},[124],{"categories":1494},[],{"categories":1496},[124],{"categories":1498},[],{"categories":1500},[],{"categories":1502},[124],{"categories":1504},[124],{"categories":1506},[124],{"categories":1508},[127],{"categories":1510},[124],{"categories":1512},[],{"categories":1514},[169],{"categories":1516},[127],{"categories":1518},[],{"categories":1520},[124],{"categories":1522},[145],{"categories":1524},[],{"categories":1526},[166],{"categories":1528},[438],{"categories":1530},[145],{"categories":1532},[176],{"categories":1534},[176],{"categories":1536},[145],{"categories":1538},[145],{"categories":1540},[438],{"categories":1542},[],{"categories":1544},[145],{"categories":1546},[124],{"categories":1548},[118],{"categories":1550},[145],{"categories":1552},[],{"categories":1554},[169],{"categories":1556},[145],{"categories":1558},[176],{"categories":1560},[145],{"categories":1562},[438],{"categories":1564},[124],{"categories":1566},[124],{"categories":1568},[],{"categories":1570},[121],{"categories":1572},[],{"categories":1574},[],{"categories":1576},[124],{"categories":1578},[124],{"categories":1580},[124],{"categories":1582},[124],{"categories":1584},[],{"categories":1586},[169],{"categories":1588},[118],{"categories":1590},[],{"categories":1592},[124],{"categories":1594},[124],{"categories":1596},[438],{"categories":1598},[438],{"categories":1600},[],{"categories":1602},[127],{"categories":1604},[145],{"categories":1606},[145],{"categories":1608},[124],{"categories":1610},[127],{"categories":1612},[],{"categories":1614},[166],{"categories":1616},[124],{"categories":1618},[124],{"categories":1620},[],{"categories":1622},[],{"categories":1624},[438],{"categories":1626},[124],{"categories":1628},[176],{"categories":1630},[121],{"categories":1632},[124],{"categories":1634},[],{"categories":1636},[127],{"categories":1638},[118],{"categories":1640},[118],{"categories":1642},[],{"categories":1644},[124],{"categories":1646},[166],{"categories":1648},[127],{"categories":1650},[],{"categories":1652},[124],{"categories":1654},[124],{"categories":1656},[127],{"categories":1658},[],{"categories":1660},[127],{"categories":1662},[176],{"categories":1664},[],{"categories":1666},[124],{"categories":1668},[],{"categories":1670},[124],{"categories":1672},[],{"categories":1674},[124],{"categories":1676},[124],{"categories":1678},[],{"categories":1680},[124],{"categories":1682},[145],{"categories":1684},[124],{"categories":1686},[124],{"categories":1688},[118],{"categories":1690},[124],{"categories":1692},[145],{"categories":1694},[127],{"categories":1696},[],{"categories":1698},[124],{"categories":1700},[183],{"categories":1702},[],{"categories":1704},[],{"categories":1706},[],{"categories":1708},[118],{"categories":1710},[145],{"categories":1712},[127],{"categories":1714},[124],{"categories":1716},[166],{"categories":1718},[127],{"categories":1720},[],{"categories":1722},[127],{"categories":1724},[],{"categories":1726},[124],{"categories":1728},[127],{"categories":1730},[124],{"categories":1732},[],{"categories":1734},[124],{"categories":1736},[124],{"categories":1738},[145],{"categories":1740},[166],{"categories":1742},[127],{"categories":1744},[166],{"categories":1746},[121],{"categories":1748},[],{"categories":1750},[],{"categories":1752},[124],{"categories":1754},[118],{"categories":1756},[145],{"categories":1758},[],{"categories":1760},[],{"categories":1762},[176],{"categories":1764},[166],{"categories":1766},[],{"categories":1768},[124],{"categories":1770},[],{"categories":1772},[183],{"categories":1774},[124],{"categories":1776},[438],{"categories":1778},[176],{"categories":1780},[],{"categories":1782},[127],{"categories":1784},[124],{"categories":1786},[127],{"categories":1788},[127],{"categories":1790},[124],{"categories":1792},[],{"categories":1794},[118],{"categories":1796},[124],{"categories":1798},[121],{"categories":1800},[176],{"categories":1802},[166],{"categories":1804},[],{"categories":1806},[],{"categories":1808},[],{"categories":1810},[127],{"categories":1812},[166],{"categories":1814},[145],{"categories":1816},[124],{"categories":1818},[145],{"categories":1820},[166],{"categories":1822},[],{"categories":1824},[166],{"categories":1826},[145],{"categories":1828},[121],{"categories":1830},[124],{"categories":1832},[145],{"categories":1834},[183],{"categories":1836},[],{"categories":1838},[],{"categories":1840},[169],{"categories":1842},[124,176],{"categories":1844},[145],{"categories":1846},[124],{"categories":1848},[127],{"categories":1850},[127],{"categories":1852},[124],{"categories":1854},[],{"categories":1856},[176],{"categories":1858},[124],{"categories":1860},[169],{"categories":1862},[127],{"categories":1864},[183],{"categories":1866},[438],{"categories":1868},[],{"categories":1870},[118],{"categories":1872},[127],{"categories":1874},[127],{"categories":1876},[176],{"categories":1878},[124],{"categories":1880},[124],{"categories":1882},[],{"categories":1884},[],{"categories":1886},[],{"categories":1888},[438],{"categories":1890},[145],{"categories":1892},[124],{"categories":1894},[124],{"categories":1896},[124],{"categories":1898},[],{"categories":1900},[169],{"categories":1902},[121],{"categories":1904},[],{"categories":1906},[127],{"categories":1908},[438],{"categories":1910},[],{"categories":1912},[166],{"categories":1914},[166],{"categories":1916},[],{"categories":1918},[176],{"categories":1920},[166],{"categories":1922},[124],{"categories":1924},[],{"categories":1926},[145],{"categories":1928},[124],{"categories":1930},[166],{"categories":1932},[127],{"categories":1934},[145],{"categories":1936},[],{"categories":1938},[127],{"categories":1940},[166],{"categories":1942},[124],{"categories":1944},[],{"categories":1946},[124],{"categories":1948},[124],{"categories":1950},[438],{"categories":1952},[145],{"categories":1954},[169],{"categories":1956},[169],{"categories":1958},[],{"categories":1960},[],{"categories":1962},[],{"categories":1964},[127],{"categories":1966},[176],{"categories":1968},[176],{"categories":1970},[],{"categories":1972},[],{"categories":1974},[124],{"categories":1976},[],{"categories":1978},[127],{"categories":1980},[124],{"categories":1982},[],{"categories":1984},[124],{"categories":1986},[121],{"categories":1988},[124],{"categories":1990},[183],{"categories":1992},[127],{"categories":1994},[124],{"categories":1996},[176],{"categories":1998},[145],{"categories":2000},[127],{"categories":2002},[],{"categories":2004},[145],{"categories":2006},[127],{"categories":2008},[127],{"categories":2010},[],{"categories":2012},[121],{"categories":2014},[127],{"categories":2016},[],{"categories":2018},[124],{"categories":2020},[118],{"categories":2022},[145],{"categories":2024},[438],{"categories":2026},[127],{"categories":2028},[127],{"categories":2030},[118],{"categories":2032},[124],{"categories":2034},[],{"categories":2036},[],{"categories":2038},[166],{"categories":2040},[124,121],{"categories":2042},[],{"categories":2044},[118],{"categories":2046},[169],{"categories":2048},[124],{"categories":2050},[176],{"categories":2052},[124],{"categories":2054},[127],{"categories":2056},[124],{"categories":2058},[124],{"categories":2060},[145],{"categories":2062},[127],{"categories":2064},[],{"categories":2066},[],{"categories":2068},[127],{"categories":2070},[124],{"categories":2072},[438],{"categories":2074},[],{"categories":2076},[124],{"categories":2078},[127],{"categories":2080},[],{"categories":2082},[124],{"categories":2084},[183],{"categories":2086},[169],{"categories":2088},[127],{"categories":2090},[124],{"categories":2092},[438],{"categories":2094},[],{"categories":2096},[124],{"categories":2098},[183],{"categories":2100},[166],{"categories":2102},[124],{"categories":2104},[],{"categories":2106},[183],{"categories":2108},[145],{"categories":2110},[124],{"categories":2112},[124],{"categories":2114},[118],{"categories":2116},[],{"categories":2118},[],{"categories":2120},[166],{"categories":2122},[124],{"categories":2124},[169],{"categories":2126},[183],{"categories":2128},[183],{"categories":2130},[145],{"categories":2132},[],{"categories":2134},[],{"categories":2136},[124],{"categories":2138},[],{"categories":2140},[124,176],{"categories":2142},[145],{"categories":2144},[127],{"categories":2146},[176],{"categories":2148},[124],{"categories":2150},[118],{"categories":2152},[],{"categories":2154},[],{"categories":2156},[118],{"categories":2158},[183],{"categories":2160},[124],{"categories":2162},[],{"categories":2164},[166,124],{"categories":2166},[438],{"categories":2168},[118],{"categories":2170},[],{"categories":2172},[121],{"categories":2174},[121],{"categories":2176},[124],{"categories":2178},[176],{"categories":2180},[127],{"categories":2182},[145],{"categories":2184},[183],{"categories":2186},[166],{"categories":2188},[124],{"categories":2190},[124],{"categories":2192},[124],{"categories":2194},[118],{"categories":2196},[124],{"categories":2198},[127],{"categories":2200},[145],{"categories":2202},[],{"categories":2204},[],{"categories":2206},[169],{"categories":2208},[176],{"categories":2210},[124],{"categories":2212},[166],{"categories":2214},[169],{"categories":2216},[124],{"categories":2218},[124],{"categories":2220},[127],{"categories":2222},[127],{"categories":2224},[124,121],{"categories":2226},[],{"categories":2228},[166],{"categories":2230},[],{"categories":2232},[124],{"categories":2234},[145],{"categories":2236},[118],{"categories":2238},[118],{"categories":2240},[127],{"categories":2242},[124],{"categories":2244},[121],{"categories":2246},[176],{"categories":2248},[183],{"categories":2250},[],{"categories":2252},[145],{"categories":2254},[124],{"categories":2256},[124],{"categories":2258},[145],{"categories":2260},[176],{"categories":2262},[124],{"categories":2264},[127],{"categories":2266},[145],{"categories":2268},[124],{"categories":2270},[166],{"categories":2272},[124],{"categories":2274},[124],{"categories":2276},[438],{"categories":2278},[130],{"categories":2280},[127],{"categories":2282},[124],{"categories":2284},[145],{"categories":2286},[127],{"categories":2288},[183],{"categories":2290},[124],{"categories":2292},[],{"categories":2294},[124],{"categories":2296},[],{"categories":2298},[],{"categories":2300},[],{"categories":2302},[121],{"categories":2304},[124],{"categories":2306},[127],{"categories":2308},[145],{"categories":2310},[145],{"categories":2312},[145],{"categories":2314},[145],{"categories":2316},[],{"categories":2318},[118],{"categories":2320},[127],{"categories":2322},[145],{"categories":2324},[118],{"categories":2326},[127],{"categories":2328},[124],{"categories":2330},[124,127],{"categories":2332},[127],{"categories":2334},[438],{"categories":2336},[145],{"categories":2338},[145],{"categories":2340},[127],{"categories":2342},[124],{"categories":2344},[],{"categories":2346},[145],{"categories":2348},[183],{"categories":2350},[118],{"categories":2352},[124],{"categories":2354},[124],{"categories":2356},[],{"categories":2358},[176],{"categories":2360},[],{"categories":2362},[118],{"categories":2364},[127],{"categories":2366},[145],{"categories":2368},[124],{"categories":2370},[145],{"categories":2372},[118],{"categories":2374},[145],{"categories":2376},[145],{"categories":2378},[],{"categories":2380},[121],{"categories":2382},[127],{"categories":2384},[145],{"categories":2386},[145],{"categories":2388},[145],{"categories":2390},[145],{"categories":2392},[145],{"categories":2394},[145],{"categories":2396},[145],{"categories":2398},[145],{"categories":2400},[145],{"categories":2402},[145],{"categories":2404},[169],{"categories":2406},[118],{"categories":2408},[124],{"categories":2410},[124],{"categories":2412},[],{"categories":2414},[124,118],{"categories":2416},[],{"categories":2418},[127],{"categories":2420},[145],{"categories":2422},[127],{"categories":2424},[124],{"categories":2426},[124],{"categories":2428},[124],{"categories":2430},[124],{"categories":2432},[124],{"categories":2434},[127],{"categories":2436},[121],{"categories":2438},[166],{"categories":2440},[145],{"categories":2442},[124],{"categories":2444},[],{"categories":2446},[],{"categories":2448},[127],{"categories":2450},[166],{"categories":2452},[124],{"categories":2454},[],{"categories":2456},[],{"categories":2458},[183],{"categories":2460},[124],{"categories":2462},[],{"categories":2464},[],{"categories":2466},[118],{"categories":2468},[121],{"categories":2470},[124],{"categories":2472},[121],{"categories":2474},[166],{"categories":2476},[],{"categories":2478},[145],{"categories":2480},[],{"categories":2482},[166],{"categories":2484},[124],{"categories":2486},[183],{"categories":2488},[],{"categories":2490},[183],{"categories":2492},[],{"categories":2494},[],{"categories":2496},[127],{"categories":2498},[],{"categories":2500},[121],{"categories":2502},[118],{"categories":2504},[166],{"categories":2506},[176],{"categories":2508},[],{"categories":2510},[],{"categories":2512},[124],{"categories":2514},[118],{"categories":2516},[183],{"categories":2518},[],{"categories":2520},[127],{"categories":2522},[127],{"categories":2524},[145],{"categories":2526},[124],{"categories":2528},[127],{"categories":2530},[124],{"categories":2532},[127],{"categories":2534},[124],{"categories":2536},[130],{"categories":2538},[145],{"categories":2540},[],{"categories":2542},[183],{"categories":2544},[176],{"categories":2546},[127],{"categories":2548},[],{"categories":2550},[124],{"categories":2552},[127],{"categories":2554},[121],{"categories":2556},[118],{"categories":2558},[124],{"categories":2560},[166],{"categories":2562},[176],{"categories":2564},[176],{"categories":2566},[124],{"categories":2568},[169],{"categories":2570},[124],{"categories":2572},[127],{"categories":2574},[121],{"categories":2576},[127],{"categories":2578},[124],{"categories":2580},[124],{"categories":2582},[127],{"categories":2584},[145],{"categories":2586},[],{"categories":2588},[118],{"categories":2590},[124],{"categories":2592},[127],{"categories":2594},[124],{"categories":2596},[124],{"categories":2598},[],{"categories":2600},[166],{"categories":2602},[121],{"categories":2604},[145],{"categories":2606},[124],{"categories":2608},[124],{"categories":2610},[166],{"categories":2612},[183],{"categories":2614},[169],{"categories":2616},[124],{"categories":2618},[145],{"categories":2620},[124],{"categories":2622},[127],{"categories":2624},[438],{"categories":2626},[124],{"categories":2628},[127],{"categories":2630},[169],{"categories":2632},[],{"categories":2634},[127],{"categories":2636},[176],{"categories":2638},[166],{"categories":2640},[124],{"categories":2642},[118],{"categories":2644},[121],{"categories":2646},[176],{"categories":2648},[],{"categories":2650},[127],{"categories":2652},[124],{"categories":2654},[],{"categories":2656},[145],{"categories":2658},[],{"categories":2660},[145],{"categories":2662},[124],{"categories":2664},[127],{"categories":2666},[127],{"categories":2668},[127],{"categories":2670},[],{"categories":2672},[],{"categories":2674},[124],{"categories":2676},[124],{"categories":2678},[],{"categories":2680},[166],{"categories":2682},[127],{"categories":2684},[183],{"categories":2686},[118],{"categories":2688},[],{"categories":2690},[],{"categories":2692},[145],{"categories":2694},[176],{"categories":2696},[124],{"categories":2698},[124],{"categories":2700},[124],{"categories":2702},[176],{"categories":2704},[145],{"categories":2706},[166],{"categories":2708},[124],{"categories":2710},[124],{"categories":2712},[124],{"categories":2714},[145],{"categories":2716},[124],{"categories":2718},[145],{"categories":2720},[127],{"categories":2722},[127],{"categories":2724},[176],{"categories":2726},[127],{"categories":2728},[124],{"categories":2730},[176],{"categories":2732},[166],{"categories":2734},[],{"categories":2736},[127],{"categories":2738},[],{"categories":2740},[],{"categories":2742},[121],{"categories":2744},[124],{"categories":2746},[127],{"categories":2748},[118],{"categories":2750},[127],{"categories":2752},[183],{"categories":2754},[],{"categories":2756},[127],{"categories":2758},[],{"categories":2760},[118],{"categories":2762},[127],{"categories":2764},[],{"categories":2766},[127],{"categories":2768},[124],{"categories":2770},[145],{"categories":2772},[124],{"categories":2774},[127],{"categories":2776},[145],{"categories":2778},[127],{"categories":2780},[176],{"categories":2782},[166],{"categories":2784},[118],{"categories":2786},[],{"categories":2788},[127],{"categories":2790},[166],{"categories":2792},[145],{"categories":2794},[124],{"categories":2796},[166],{"categories":2798},[118],{"categories":2800},[],{"categories":2802},[127],{"categories":2804},[127],{"categories":2806},[124],{"categories":2808},[],{"categories":2810},[127],{"categories":2812},[130],{"categories":2814},[145],{"categories":2816},[127],{"categories":2818},[121],{"categories":2820},[],{"categories":2822},[124],{"categories":2824},[130],{"categories":2826},[124],{"categories":2828},[127],{"categories":2830},[145],{"categories":2832},[118],{"categories":2834},[438],{"categories":2836},[124],{"categories":2838},[124],{"categories":2840},[124],{"categories":2842},[145],{"categories":2844},[121],{"categories":2846},[124],{"categories":2848},[166],{"categories":2850},[145],{"categories":2852},[438],{"categories":2854},[124],{"categories":2856},[],{"categories":2858},[],{"categories":2860},[438],{"categories":2862},[169],{"categories":2864},[127],{"categories":2866},[127],{"categories":2868},[145],{"categories":2870},[124],{"categories":2872},[118],{"categories":2874},[166],{"categories":2876},[127],{"categories":2878},[124],{"categories":2880},[183],{"categories":2882},[124],{"categories":2884},[127],{"categories":2886},[],{"categories":2888},[124],{"categories":2890},[124],{"categories":2892},[145],{"categories":2894},[118],{"categories":2896},[],{"categories":2898},[124],{"categories":2900},[124],{"categories":2902},[176],{"categories":2904},[166],{"categories":2906},[124,127],{"categories":2908},[183,121],{"categories":2910},[124],{"categories":2912},[],{"categories":2914},[127],{"categories":2916},[],{"categories":2918},[176],{"categories":2920},[124],{"categories":2922},[145],{"categories":2924},[],{"categories":2926},[127],{"categories":2928},[],{"categories":2930},[127],{"categories":2932},[118],{"categories":2934},[127],{"categories":2936},[124],{"categories":2938},[438],{"categories":2940},[183],{"categories":2942},[121],{"categories":2944},[121],{"categories":2946},[118],{"categories":2948},[118],{"categories":2950},[124],{"categories":2952},[127],{"categories":2954},[124],{"categories":2956},[124],{"categories":2958},[118],{"categories":2960},[124],{"categories":2962},[183],{"categories":2964},[145],{"categories":2966},[124],{"categories":2968},[127],{"categories":2970},[124],{"categories":2972},[],{"categories":2974},[176],{"categories":2976},[],{"categories":2978},[127],{"categories":2980},[118],{"categories":2982},[],{"categories":2984},[438],{"categories":2986},[124],{"categories":2988},[],{"categories":2990},[145],{"categories":2992},[127],{"categories":2994},[176],{"categories":2996},[124],{"categories":2998},[127],{"categories":3000},[176],{"categories":3002},[127],{"categories":3004},[145],{"categories":3006},[118],{"categories":3008},[145],{"categories":3010},[176],{"categories":3012},[124],{"categories":3014},[166],{"categories":3016},[124],{"categories":3018},[124],{"categories":3020},[124],{"categories":3022},[124],{"categories":3024},[127],{"categories":3026},[124],{"categories":3028},[127],{"categories":3030},[124],{"categories":3032},[118],{"categories":3034},[124],{"categories":3036},[127],{"categories":3038},[166],{"categories":3040},[118],{"categories":3042},[127],{"categories":3044},[166],{"categories":3046},[],{"categories":3048},[124],{"categories":3050},[124],{"categories":3052},[176],{"categories":3054},[],{"categories":3056},[127],{"categories":3058},[183],{"categories":3060},[124],{"categories":3062},[145],{"categories":3064},[183],{"categories":3066},[127],{"categories":3068},[121],{"categories":3070},[121],{"categories":3072},[124],{"categories":3074},[118],{"categories":3076},[],{"categories":3078},[124],{"categories":3080},[],{"categories":3082},[118],{"categories":3084},[124],{"categories":3086},[127],{"categories":3088},[127],{"categories":3090},[],{"categories":3092},[176],{"categories":3094},[176],{"categories":3096},[183],{"categories":3098},[166],{"categories":3100},[],{"categories":3102},[124],{"categories":3104},[118],{"categories":3106},[124],{"categories":3108},[176],{"categories":3110},[118],{"categories":3112},[145],{"categories":3114},[145],{"categories":3116},[],{"categories":3118},[145],{"categories":3120},[127],{"categories":3122},[166],{"categories":3124},[169],{"categories":3126},[124],{"categories":3128},[],{"categories":3130},[145],{"categories":3132},[176],{"categories":3134},[121],{"categories":3136},[124],{"categories":3138},[118],{"categories":3140},[438],{"categories":3142},[118],{"categories":3144},[],{"categories":3146},[],{"categories":3148},[145],{"categories":3150},[],{"categories":3152},[127],{"categories":3154},[127],{"categories":3156},[127],{"categories":3158},[],{"categories":3160},[124],{"categories":3162},[],{"categories":3164},[145],{"categories":3166},[118],{"categories":3168},[166],{"categories":3170},[124],{"categories":3172},[145],{"categories":3174},[145],{"categories":3176},[],{"categories":3178},[145],{"categories":3180},[118],{"categories":3182},[124],{"categories":3184},[],{"categories":3186},[127],{"categories":3188},[127],{"categories":3190},[118],{"categories":3192},[],{"categories":3194},[],{"categories":3196},[],{"categories":3198},[166],{"categories":3200},[127],{"categories":3202},[124],{"categories":3204},[],{"categories":3206},[],{"categories":3208},[],{"categories":3210},[166],{"categories":3212},[],{"categories":3214},[118],{"categories":3216},[],{"categories":3218},[],{"categories":3220},[166],{"categories":3222},[124],{"categories":3224},[145],{"categories":3226},[],{"categories":3228},[183],{"categories":3230},[145],{"categories":3232},[183],{"categories":3234},[124],{"categories":3236},[],{"categories":3238},[],{"categories":3240},[127],{"categories":3242},[],{"categories":3244},[],{"categories":3246},[127],{"categories":3248},[124],{"categories":3250},[],{"categories":3252},[127],{"categories":3254},[145],{"categories":3256},[183],{"categories":3258},[169],{"categories":3260},[127],{"categories":3262},[127],{"categories":3264},[],{"categories":3266},[],{"categories":3268},[],{"categories":3270},[145],{"categories":3272},[],{"categories":3274},[],{"categories":3276},[166],{"categories":3278},[118],{"categories":3280},[],{"categories":3282},[121],{"categories":3284},[183],{"categories":3286},[124],{"categories":3288},[176],{"categories":3290},[118],{"categories":3292},[169],{"categories":3294},[121],{"categories":3296},[176],{"categories":3298},[],{"categories":3300},[],{"categories":3302},[127],{"categories":3304},[118],{"categories":3306},[166],{"categories":3308},[118],{"categories":3310},[127],{"categories":3312},[438],{"categories":3314},[127],{"categories":3316},[],{"categories":3318},[124],{"categories":3320},[145],{"categories":3322},[176],{"categories":3324},[],{"categories":3326},[166],{"categories":3328},[145],{"categories":3330},[118],{"categories":3332},[127],{"categories":3334},[124],{"categories":3336},[121],{"categories":3338},[127,438],{"categories":3340},[127],{"categories":3342},[176],{"categories":3344},[124],{"categories":3346},[169],{"categories":3348},[183],{"categories":3350},[127],{"categories":3352},[],{"categories":3354},[127],{"categories":3356},[124],{"categories":3358},[121],{"categories":3360},[],{"categories":3362},[],{"categories":3364},[124],{"categories":3366},[169],{"categories":3368},[124],{"categories":3370},[],{"categories":3372},[145],{"categories":3374},[],{"categories":3376},[145],{"categories":3378},[176],{"categories":3380},[127],{"categories":3382},[124],{"categories":3384},[183],{"categories":3386},[176],{"categories":3388},[],{"categories":3390},[145],{"categories":3392},[124],{"categories":3394},[],{"categories":3396},[124],{"categories":3398},[127],{"categories":3400},[124],{"categories":3402},[127],{"categories":3404},[124],{"categories":3406},[124],{"categories":3408},[124],{"categories":3410},[124],{"categories":3412},[121],{"categories":3414},[],{"categories":3416},[130],{"categories":3418},[145],{"categories":3420},[124],{"categories":3422},[],{"categories":3424},[176],{"categories":3426},[124],{"categories":3428},[124],{"categories":3430},[127],{"categories":3432},[145],{"categories":3434},[124],{"categories":3436},[124],{"categories":3438},[121],{"categories":3440},[127],{"categories":3442},[166],{"categories":3444},[],{"categories":3446},[169],{"categories":3448},[124],{"categories":3450},[],{"categories":3452},[145],{"categories":3454},[183],{"categories":3456},[],{"categories":3458},[],{"categories":3460},[145],{"categories":3462},[145],{"categories":3464},[183],{"categories":3466},[118],{"categories":3468},[127],{"categories":3470},[127],{"categories":3472},[124],{"categories":3474},[121],{"categories":3476},[],{"categories":3478},[],{"categories":3480},[145],{"categories":3482},[169],{"categories":3484},[176],{"categories":3486},[127],{"categories":3488},[166],{"categories":3490},[169],{"categories":3492},[169],{"categories":3494},[],{"categories":3496},[145],{"categories":3498},[124],{"categories":3500},[124],{"categories":3502},[176],{"categories":3504},[],{"categories":3506},[145],{"categories":3508},[145],{"categories":3510},[145],{"categories":3512},[],{"categories":3514},[127],{"categories":3516},[124],{"categories":3518},[],{"categories":3520},[118],{"categories":3522},[121],{"categories":3524},[],{"categories":3526},[124],{"categories":3528},[124],{"categories":3530},[],{"categories":3532},[176],{"categories":3534},[],{"categories":3536},[],{"categories":3538},[],{"categories":3540},[],{"categories":3542},[124],{"categories":3544},[145],{"categories":3546},[],{"categories":3548},[],{"categories":3550},[124],{"categories":3552},[124],{"categories":3554},[124],{"categories":3556},[169],{"categories":3558},[124],{"categories":3560},[169],{"categories":3562},[],{"categories":3564},[169],{"categories":3566},[169],{"categories":3568},[438],{"categories":3570},[127],{"categories":3572},[176],{"categories":3574},[],{"categories":3576},[],{"categories":3578},[169],{"categories":3580},[176],{"categories":3582},[176],{"categories":3584},[176],{"categories":3586},[],{"categories":3588},[118],{"categories":3590},[176],{"categories":3592},[176],{"categories":3594},[118],{"categories":3596},[176],{"categories":3598},[121],{"categories":3600},[176],{"categories":3602},[176],{"categories":3604},[176],{"categories":3606},[169],{"categories":3608},[145],{"categories":3610},[145],{"categories":3612},[124],{"categories":3614},[176],{"categories":3616},[169],{"categories":3618},[438],{"categories":3620},[169],{"categories":3622},[169],{"categories":3624},[169],{"categories":3626},[],{"categories":3628},[121],{"categories":3630},[],{"categories":3632},[438],{"categories":3634},[176],{"categories":3636},[176],{"categories":3638},[176],{"categories":3640},[127],{"categories":3642},[145,121],{"categories":3644},[169],{"categories":3646},[],{"categories":3648},[],{"categories":3650},[169],{"categories":3652},[],{"categories":3654},[169],{"categories":3656},[145],{"categories":3658},[127],{"categories":3660},[],{"categories":3662},[176],{"categories":3664},[124],{"categories":3666},[166],{"categories":3668},[],{"categories":3670},[124],{"categories":3672},[],{"categories":3674},[145],{"categories":3676},[118],{"categories":3678},[169],{"categories":3680},[],{"categories":3682},[176],{"categories":3684},[145],[3686,3918,4004,4089],{"id":3687,"title":3688,"ai":3689,"body":3694,"categories":3884,"created_at":79,"date_modified":79,"description":72,"extension":80,"faq":79,"featured":81,"kicker_label":79,"meta":3885,"navigation":97,"path":3905,"published_at":3906,"question":79,"scraped_at":3907,"seo":3908,"sitemap":3909,"source_id":3910,"source_name":3911,"source_type":104,"source_url":3912,"stem":3913,"tags":3914,"thumbnail_url":79,"tldr":3915,"tweet":79,"unknown_tags":3916,"__hash__":3917},"summaries\u002Fsummaries\u002F4311686432e3e5ff-tiny-llms-and-on-device-agents-via-litert-lm-on-ed-summary.md","Tiny LLMs and On-Device Agents via LiteRT-LM on Edge Hardware",{"provider":7,"model":8,"input_tokens":3690,"output_tokens":3691,"processing_time_ms":3692,"cost_usd":3693},8771,2602,22997,0.0030327,{"type":14,"value":3695,"toc":3876},[3696,3700,3703,3706,3709,3713,3721,3727,3730,3738,3742,3745,3748,3815,3818,3821,3825,3828,3831,3834,3838,3841,3844,3848],[17,3697,3699],{"id":3698},"edge-ai-benefits-drive-on-device-llms","Edge AI Benefits Drive On-Device LLMs",[22,3701,3702],{},"Running LLMs on edge devices solves key constraints: ultra-low latency for in-loop UX like live voice translation (impossible via cloud), full privacy in messaging apps, offline capability, and cost savings on laptops. Cormac Brick, Google AI Edge tech lead, emphasizes these over cloud alternatives, drawing from 10 years optimizing hardware from Raspberry Pi to NPUs. Tradeoffs include RAM limits (e.g., 2-4GB for viable models) and hardware variability, pushing optimizations like memory-mapped per-layer embeddings to keep effective params low.",[22,3704,3705],{},"\"There's a lot of benefits to running on the edge. There's latency or UX improvements for some really sensitive in-the-loop things like live voice translation.\" — Cormac Brick, highlighting why Pixel's on-device translation beats cloud latency.",[22,3707,3708],{},"Google's stack—LiteRT (ex-TensorFlow Lite), MediaPipe, LiteRT-LM—ships in Photos, YouTube Shorts effects, and Android system services. One .tflite file deploys cross-platform (Android\u002FiOS\u002FMac\u002FLinux\u002FWindows\u002FWeb\u002FIoT) on CPU\u002FGPU; NPUs need separate compilation. This enables broad reach beyond premium devices.",[17,3710,3712],{"id":3711},"system-genai-vs-in-app-tiny-llms-deployment-patterns","System GenAI vs. In-App Tiny LLMs: Deployment Patterns",[22,3714,3715,3716,3720],{},"Two trends emerge: ",[3717,3718,3719],"strong",{},"system-level GenAI"," integrates 2-5B param models into OS (Android AI Core, Apple Intelligence) for broad APIs like summarization\u002Fprompting, pre-loaded on premium devices. Customization via prompting or skills; no app downloads needed.",[22,3722,3723,3726],{},[3717,3724,3725],{},"In-app GenAI"," uses tiny LLMs (TLMs, 100-500M params) bundled with apps\u002Fwebpages for wider device compatibility. Fine-tuning is essential below 500M params for production reliability on tasks like summarization, transcription, voice-to-function (e.g., Function Gemma at 270M params hits 85-90% on 10 Android functions). Prompting alone fails for tiny models; fine-tuning yields \"really reliable performance.\"",[22,3728,3729],{},"Decision chain: System for foundation tasks (leverage OS investment); in-app for custom, task-specific reliability. Tradeoff: System limits to premium hardware; tiny models sacrifice generality but gain deployability.",[22,3731,3732,3733,3737],{},"\"For the really really tiny models certainly less than 500 ",[3734,3735,3736],"span",{},"million parameters"," you need to fine-tune to get production level reliability.\" — Brick on why prompting isn't enough for edge-scale models.",[17,3739,3741],{"id":3740},"gemma-2b4b-edge-optimized-for-agents-and-multimodality","Gemma 2B\u002F4B: Edge-Optimized for Agents and Multimodality",[22,3743,3744],{},"Gemma 2 (E2B: 2B effective params; E4B: 4B) targets edge with RAM efficiency via partial embedding loads (hundreds of bytes per token). Multimodal (audio\u002Fimage\u002Ftext for small sizes); built-in function calling + thinking unlocks on-device agents. Apache 2.0 license broadens use.",[22,3746,3747],{},"Performance (snapshot, ongoing optimizations with Qualcomm\u002FIntel\u002FRaspberry Pi):",[3749,3750,3751,3767],"table",{},[3752,3753,3754],"thead",{},[3755,3756,3757,3761,3764],"tr",{},[3758,3759,3760],"th",{},"Device",[3758,3762,3763],{},"Gemma 2B Prefill\u002FDecode (tok\u002Fs)",[3758,3765,3766],{},"Gemma 4B Prefill\u002FDecode (tok\u002Fs)",[3768,3769,3770,3782,3793,3804],"tbody",{},[3755,3771,3772,3776,3779],{},[3773,3774,3775],"td",{},"High-end Android (GPU)",[3773,3777,3778],{},"2000+\u002F1000+",[3773,3780,3781],{},"~half",[3755,3783,3784,3787,3790],{},[3773,3785,3786],{},"MacBook",[3773,3788,3789],{},"1000s",[3773,3791,3792],{},"Proportional",[3755,3794,3795,3798,3801],{},[3773,3796,3797],{},"Raspberry Pi 5",[3773,3799,3800],{},"20\u002F133",[3773,3802,3803],{},"N\u002FA",[3755,3805,3806,3809,3812],{},[3773,3807,3808],{},"Qualcomm IoT NPU",[3773,3810,3811],{},"High (NPU boost)",[3773,3813,3814],{},"High",[22,3816,3817],{},"E2B\u002F4B on AI Core roadmap for Android integration. Larger Gemma for laptops (32GB RAM).",[22,3819,3820],{},"\"One of the big step ups... was they've kind of built in function calling which is excellent and they also have built-in thinking. So that combination... unlocks our ability to now do skills on device.\" — Brick on Gemma's agent enablers.",[17,3822,3824],{"id":3823},"progressive-skills-token-efficient-on-device-agents","Progressive Skills: Token-Efficient On-Device Agents",[22,3826,3827],{},"Google AI Gallery app demos agent skills: mood journaling (log\u002Fanalyze trends via voice), calendar checks, Wikipedia queries, music synthesis from images. No fine-tuning; skills as on-demand JS snippets with one-line descriptions.",[22,3829,3830],{},"Mechanism: Progressive disclosure—model sees skill summaries first, loads details (functions) only if relevant via a \"load skill\" meta-function. Cuts context bloat, boosts reliability on lightweight models (poor at long contexts). Patterns: knowledge augmentation (Wikipedia), interactive UI (flashcards), web services (weather\u002Fmaps\u002Fmusic).",[22,3832,3833],{},"\"The way we've built the skills is there's a kind of one-line description... if it thinks that sounds interesting, then it asks for more... This is particularly important for token efficiency and frankly reliability on edge models.\" — Brick explaining conditional depth over full MCP descriptions.",[17,3835,3837],{"id":3836},"tiny-model-workflow-fine-tune-and-deploy","Tiny Model Workflow: Fine-Tune and Deploy",[22,3839,3840],{},"For TLMs: Fine-tune Gemma-based models (e.g., 100-500M) on task data, quantize, deploy via LiteRT-LM. Example app (team-built): Real-world tiny LLM use, voice-to-action. Cross-platform speed via hardware accel (GPU\u002FNPU).",[22,3842,3843],{},"Tradeoffs: Tiny = task-specific excellence but no generality; needs fine-tuning. Results: Voice-to-function at 85-90% on small models, deployable everywhere.",[17,3845,3847],{"id":3846},"key-takeaways","Key Takeaways",[3849,3850,3851,3855,3858,3861,3864,3867,3870,3873],"ul",{},[3852,3853,3854],"li",{},"Prioritize edge for latency\u002Fprivacy\u002Foffline\u002Fcost; use LiteRT-LM for cross-platform .tflite deployment (CPU\u002FGPU standard, NPU compiled).",[3852,3856,3857],{},"Choose system GenAI (2-5B params via OS APIs) for foundation tasks on premium devices; in-app TLMs (100-500M) for custom tasks with fine-tuning.",[3852,3859,3860],{},"Gemma 2B\u002F4B: 2-4GB RAM effective, multimodal, agent-ready; expect 100-2000+ tok\u002Fs depending on hardware.",[3852,3862,3863],{},"Build skills progressively: One-line summaries → on-demand JS loads for token efficiency and dynamic tools.",[3852,3865,3866],{},"Fine-tune tiny models below 500M params for 85-90% reliability on voice\u002Faction tasks; avoid prompting alone.",[3852,3868,3869],{},"Optimize embeddings (memory-map PLE) to fit RAM constraints; track partners like Qualcomm for NPU gains.",[3852,3871,3872],{},"Test on real hardware: Raspberry Pi 133 tok\u002Fs decode viable for simple analysis; high-end phones hit production speeds.",[3852,3874,3875],{},"Extend models low-code: Wikipedia\u002Fmaps\u002Fmusic skills turn static LLMs into fresh-knowledge agents.",{"title":72,"searchDepth":73,"depth":73,"links":3877},[3878,3879,3880,3881,3882,3883],{"id":3698,"depth":73,"text":3699},{"id":3711,"depth":73,"text":3712},{"id":3740,"depth":73,"text":3741},{"id":3823,"depth":73,"text":3824},{"id":3836,"depth":73,"text":3837},{"id":3846,"depth":73,"text":3847},[124],{"content_references":3886,"triage":3902},[3887,3890,3892,3894,3897,3899],{"type":3888,"title":3889,"context":88},"tool","LiteRT-LM",{"type":3888,"title":3891,"context":88},"MediaPipe",{"type":3888,"title":3893,"context":88},"LiteRT",{"type":3888,"title":3895,"author":3896,"context":88},"Gemma 2B","Google DeepMind",{"type":3888,"title":3898,"context":88},"Google AI Gallery",{"type":3900,"title":3901,"context":88},"event","NeurIPS 2016",{"relevance":93,"novelty":94,"quality":94,"actionability":94,"composite":3903,"reasoning":3904},4.35,"Category: AI & LLMs. The article discusses the practical implementation of LLMs on edge devices, addressing specific pain points like latency and privacy, which are crucial for product builders. It provides insights into deployment patterns and performance metrics that can guide developers in choosing the right model for their applications.","\u002Fsummaries\u002F4311686432e3e5ff-tiny-llms-and-on-device-agents-via-litert-lm-on-ed-summary","2026-05-03 22:00:06","2026-05-04 16:07:29",{"title":3688,"description":72},{"loc":3905},"916b0f9e88910f87","AI Engineer","https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=BKWpYIWvAo4","summaries\u002F4311686432e3e5ff-tiny-llms-and-on-device-agents-via-litert-lm-on-ed-summary",[109,108,110,111],"LiteRT-LM runs Gemma 2B\u002F4B models at 1000+ tokens\u002Fsec on phones and delivers agent skills with function calling, while tiny 100-500M param models excel in fine-tuned in-app tasks like voice-to-action at 85-90% reliability.",[],"7YIkCr0_vZOvIvcSbF_z5cixaYozDoCb3U_wlYAfm8E",{"id":3919,"title":3920,"ai":3921,"body":3926,"categories":3971,"created_at":79,"date_modified":79,"description":72,"extension":80,"faq":79,"featured":81,"kicker_label":79,"meta":3972,"navigation":97,"path":3991,"published_at":3992,"question":79,"scraped_at":3993,"seo":3994,"sitemap":3995,"source_id":3996,"source_name":3997,"source_type":104,"source_url":3998,"stem":3999,"tags":4000,"thumbnail_url":79,"tldr":4001,"tweet":79,"unknown_tags":4002,"__hash__":4003},"summaries\u002Fsummaries\u002F11ccc96d3ca22d5b-gemma-chat-offline-vibe-coding-with-gemma-4-on-mac-summary.md","Gemma Chat: Offline Vibe Coding with Gemma 4 on Mac",{"provider":7,"model":8,"input_tokens":3922,"output_tokens":3923,"processing_time_ms":3924,"cost_usd":3925},6334,1865,16673,0.0021768,{"type":14,"value":3927,"toc":3966},[3928,3932,3935,3938,3942,3945,3948,3952,3963],[17,3929,3931],{"id":3930},"build-and-iterate-small-apps-offline-with-privacy","Build and Iterate Small Apps Offline with Privacy",[22,3933,3934],{},"Use Gemma Chat's Build Mode to prompt Gemma 4 for small web apps like landing pages, Pomodoro timers, dashboards, or games (e.g., Chrome Dino clone with keyboard controls). The agent creates, edits, reads files in a sandbox workspace, runs bash commands, and updates a live preview in real-time—even streaming partial file writes every few hundred milliseconds for a dynamic build experience. Switch to Chat Mode for general assistance with tools like calculations, web search, or URL fetching (online only). Voice input via local Whisper speech-to-text in the browser keeps everything on-device, ensuring prompts, code, and files stay private without cloud transmission.",[22,3936,3937],{},"This local-first setup trades cloud model power for zero API costs and full control: download models once (e.g., 3GB E4B recommended for balanced speed\u002Fcapability), then work offline on planes or private prototypes. Smaller E2B suits 8GB Macs for speed; larger MoE or 31B dense models leverage 16-32GB RAM for better reasoning on complex tasks.",[17,3939,3941],{"id":3940},"xml-tool-protocol-boosts-reliability-on-local-models","XML Tool Protocol Boosts Reliability on Local Models",[22,3943,3944],{},"Gemma Chat uses a simple XML-style protocol for tools (write file, edit file, read file, list files, run bash, open preview) instead of JSON function calling, which smaller local models handle more reliably. An MLX server streams model output to the Electron app interface, enabling agent loops where the model observes results and iterates. This powers vibe coding workflows similar to Bolt or Replit AI builders but fully local via Apple's MLX framework on Apple Silicon.",[22,3946,3947],{},"Google's Gemma 4 excels here due to its focus on agentic workflows, code generation, and local deployment—positioned by DeepMind as their strongest open family yet. Backed by Google AI Studio's Ammar Reshi (MIT-licensed repo) and promoted by the official Gemma account, it demonstrates practical local AI without benchmarks, highlighting open models' maturity for developer tools.",[17,3949,3951],{"id":3950},"setup-trade-offs-and-realistic-use-cases","Setup Trade-offs and Realistic Use Cases",[22,3953,3954,3955,3958,3959,3962],{},"Clone the GitHub repo, run ",[26,3956,3957],{},"npm install"," (Node 20+), and ",[26,3960,3961],{},"npm run dev"," (Python required); first launch downloads models and MLX. Build a DMG for distribution. Limitations include Mac-only (MLX dependency), initial internet for downloads, slower inference than cloud (e.g., Cursor\u002FClaude), and no full SaaS apps—ideal for prototypes, demos, student projects, or quick experiments where privacy or offline access matters.",[22,3964,3965],{},"Pay with hardware, not subscriptions: on Apple Silicon Macs, it replaces API bills for toy apps, letting you iterate button changes endlessly without credits. Not for production refactoring, but proves local agents are viable for real workflows, pushing open AI toward usable, permissionless coding environments.",{"title":72,"searchDepth":73,"depth":73,"links":3967},[3968,3969,3970],{"id":3930,"depth":73,"text":3931},{"id":3940,"depth":73,"text":3941},{"id":3950,"depth":73,"text":3951},[124],{"content_references":3973,"triage":3987},[3974,3978,3981,3983,3985],{"type":3888,"title":3975,"author":3976,"context":3977},"Gemma Chat","Ammar Reshi","recommended",{"type":3888,"title":3979,"author":3980,"context":88},"MLX","Apple",{"type":85,"title":3982,"author":3896,"context":88},"Gemma 4",{"type":3888,"title":3984,"context":88},"MLX-LM",{"type":3888,"title":3986,"context":88},"Whisper",{"relevance":94,"novelty":3988,"quality":94,"actionability":94,"composite":3989,"reasoning":3990},3,3.8,"Category: AI & LLMs. The article discusses using Gemma Chat for building AI-powered applications offline, addressing the audience's need for practical applications of AI tools. It provides specific examples of app types that can be built and details on the local setup, which enhances its actionability.","\u002Fsummaries\u002F11ccc96d3ca22d5b-gemma-chat-offline-vibe-coding-with-gemma-4-on-mac-summary","2026-04-30 11:26:57","2026-05-03 16:50:27",{"title":3920,"description":72},{"loc":3991},"6511d28fd46031d0","AICodeKing","https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=KnrdxmsZEqA","summaries\u002F11ccc96d3ca22d5b-gemma-chat-offline-vibe-coding-with-gemma-4-on-mac-summary",[109,110,111,108],"Gemma Chat runs Google's Gemma 4 locally on Apple Silicon Macs via MLX for private, offline app building with live previews, file editing, and agentic tools—no API keys or subscriptions needed.",[],"C-C8cqNWa2owiUSYcxL8Bu79vMFgbIsG_UdpEb182qo",{"id":4005,"title":4006,"ai":4007,"body":4012,"categories":4053,"created_at":79,"date_modified":79,"description":72,"extension":80,"faq":79,"featured":81,"kicker_label":79,"meta":4054,"navigation":97,"path":4077,"published_at":4078,"question":79,"scraped_at":4079,"seo":4080,"sitemap":4081,"source_id":4082,"source_name":3911,"source_type":104,"source_url":4083,"stem":4084,"tags":4085,"thumbnail_url":79,"tldr":4086,"tweet":79,"unknown_tags":4087,"__hash__":4088},"summaries\u002Fsummaries\u002F72b458a70b353863-gemma-4-open-models-running-ai-agents-on-device-summary.md","Gemma 4: Open Models Running AI Agents On-Device",{"provider":7,"model":8,"input_tokens":4008,"output_tokens":4009,"processing_time_ms":4010,"cost_usd":4011},7092,1839,11145,0.00183105,{"type":14,"value":4013,"toc":4048},[4014,4018,4025,4028,4032,4035,4038,4042,4045],[17,4015,4017],{"id":4016},"on-device-deployment-powers-agentic-apps","On-Device Deployment Powers Agentic Apps",[22,4019,4020,4021,4024],{},"Gemma 4 models range from 2B to 32B parameters, all fitting on consumer GPUs, laptops, phones, or even Raspberry Pi\u002FNintendo Switch via llama.cpp. The 2B\u002F4B variants run fully offline in airplane mode, generating 100 tokens\u002Fsecond for tasks like Android app coding, piano-playing agents, or parallel SVG creation (10 instances on a laptop). Use llama.cpp with the ",[26,4022,4023],{},"--override-tensor"," flag to offload per-layer embeddings to CPU\u002Fdisk, slashing GPU needs while maintaining speed. Larger 31B model maximizes raw intelligence; 26B MoE variant prioritizes low-latency inference. All support multimodal inputs (images, video, audio) for speech-to-text translation (e.g., Spanish to French) or fine-grained analysis like object detection and llama localization in photos.",[22,4026,4027],{},"LM Arena scores place Gemma 4 in the top-left quadrant: highest capability per parameter size, outperforming larger closed models in community preference for conversation\u002Fhelpfulness. Evolution from Gemma 1\u002F2\u002F3 shows consistent gains without size bloat—Gemma 3 (1B-27B) was top open model on single GPU a year ago.",[17,4029,4031],{"id":4030},"e2b-architecture-cuts-compute-for-mobile","E2B Architecture Cuts Compute for Mobile",[22,4033,4034],{},"Gemma E2B\u002FE4B (effectively 2B\u002F4B active params despite 4B-5B total) uses per-layer embeddings as lookup tables instead of matrix multiplications. Embeddings load minimally into GPU; rest stays in slower memory (CPU\u002Fdisk), ideal for mobile. This novel architecture, released last summer, enables on-device multimodality without heavy compute—e.g., Japanese text extraction from images or video understanding. Tokenizer from Gemini supports 140+ languages out-of-box, excelling in low-resource fine-tunes like Quechua or Indian languages due to multilingual design.",[22,4036,4037],{},"Apache 2.0 license (new for Gemma 4) allows full flexibility: download, fine-tune, deploy anywhere. Post-release stats: 10M base model downloads in one week, 500M total Gemma family downloads, 100k+ derived models (quantizations\u002Ffine-tunes), top Hugging Face trending.",[17,4039,4041],{"id":4040},"ecosystem-and-specialized-variants-drive-adoption","Ecosystem and Specialized Variants Drive Adoption",[22,4043,4044],{},"Integrate via Hugging Face Transformers, Unsloth, MLX, vLLM for seamless fine-tuning\u002Finference—no ecosystem switches needed. Android Studio's agent mode uses offline Gemma for code gen, boosted by Android-specific training data. Official variants: Shield Gemma for toxicity filtering in production; Med-Gemini (Gemma 3-based) for radiology\u002FX-ray analysis, further fine-tunable.",[22,4046,4047],{},"Community builds sovereign AI: AI Singapore for SE Asian languages; Sarvam (India) for official languages via government-backed models. Research highlights: DeepMind paper (Dec 2023) used Gemma 3 to propose validated cancer therapy pathways in labs. Real apps span offline Chrome extensions, finance\u002Flegal reviews, subway\u002Fplane use—prioritize open models for privacy\u002Fagentic tasks, APIs for peak intelligence. Experiment now: 1 hour playing yields insights into customizing for niches; expect massive on-device gains in 6-12 months.",{"title":72,"searchDepth":73,"depth":73,"links":4049},[4050,4051,4052],{"id":4016,"depth":73,"text":4017},{"id":4030,"depth":73,"text":4031},{"id":4040,"depth":73,"text":4041},[124],{"content_references":4055,"triage":4075},[4056,4058,4060,4062,4063,4065,4067,4069,4071],{"type":3888,"title":4057,"context":88},"llama.cpp",{"type":3888,"title":4059,"context":88},"Hugging Face Transformers",{"type":3888,"title":4061,"context":88},"Unsloth",{"type":3888,"title":3979,"context":88},{"type":3888,"title":4064,"context":88},"vLLM",{"type":3888,"title":4066,"context":88},"Android Studio",{"type":85,"title":4068,"context":88},"Shield Gemma",{"type":85,"title":4070,"context":88},"Med-Gemini",{"type":4072,"title":4073,"author":4074,"context":88},"paper","Gemma 3 for cancer therapy pathways","DeepMind researchers",{"relevance":93,"novelty":94,"quality":94,"actionability":94,"composite":3903,"reasoning":4076},"Category: AI & LLMs. The article discusses the Gemma 4 models, which are highly relevant for developers looking to integrate AI agents into their products, particularly with on-device capabilities. It provides actionable insights on deployment and architecture that can be directly applied to building AI-powered applications.","\u002Fsummaries\u002F72b458a70b353863-gemma-4-open-models-running-ai-agents-on-device-summary","2026-04-20 15:15:06","2026-04-21 15:13:10",{"title":4006,"description":72},{"loc":4077},"38b04ca9f5bb2faa","https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=_gVFUEdhCyI","summaries\u002F72b458a70b353863-gemma-4-open-models-running-ai-agents-on-device-summary",[109,111,110,108],"Gemma 4 delivers 2B-32B parameter models under Apache 2.0 that run offline on phones\u002Flaptops, handle multimodal tasks in 140+ languages, and lead LM Arena for size efficiency—enabling agentic apps like piano-playing or SVG generation without APIs.",[],"J_dWTIGTqmZj3d0-nrnLgaw8_AA0yWOMLHvZjmCIWCQ",{"id":4090,"title":4091,"ai":4092,"body":4097,"categories":4153,"created_at":79,"date_modified":79,"description":72,"extension":80,"faq":79,"featured":81,"kicker_label":79,"meta":4154,"navigation":97,"path":4173,"published_at":4174,"question":79,"scraped_at":4175,"seo":4176,"sitemap":4177,"source_id":4178,"source_name":3997,"source_type":104,"source_url":4179,"stem":4180,"tags":4181,"thumbnail_url":79,"tldr":4182,"tweet":79,"unknown_tags":4183,"__hash__":4184},"summaries\u002Fsummaries\u002F605a7bae59f3f70a-uncensored-supergemma-4-powers-local-agent-workflo-summary.md","Uncensored SuperGemma-4 Powers Local Agent Workflows",{"provider":7,"model":8,"input_tokens":4093,"output_tokens":4094,"processing_time_ms":4095,"cost_usd":4096},5669,1915,13369,0.00206865,{"type":14,"value":4098,"toc":4148},[4099,4103,4106,4109,4113,4132,4135,4139,4142,4145],[17,4100,4102],{"id":4101},"uncensored-fine-tune-enhances-gemma-4-for-practical-agents","Uncensored Fine-Tune Enhances Gemma 4 for Practical Agents",[22,4104,4105],{},"SuperGemma-4 refines Google's Gemma 4 26B A4B (instruction-tuned, 256K context, native system prompts, function calling, 3.8B active MoE params) into an uncensored variant optimized for text, coding, planning, tool-use, browser tasks, and logic—avoiding chaotic role-play while staying useful. Creator Jun Song's MLX 4-bit v2 claims QuickBench score of 95.8 (vs. 91.4 baseline) and 46.2 tokens\u002Fsecond (vs. 42.5), with gains in code, logic, Korean, and browser workflows. Use neutral embedded templates to prevent prompt drift into unwanted coding or tool modes, ensuring clean chat and agent behavior without manual chat template overrides, which can corrupt responses.",[22,4107,4108],{},"This balance makes it ideal for local power users needing permissive models that retain agent-ready architecture, outperforming stock Gemma 4 in unfiltered workflows without sacrificing reasoning.",[17,4110,4112],{"id":4111},"apple-silicon-setup-requires-24gb-ram-for-smooth-inference","Apple Silicon Setup Requires 24GB+ RAM for Smooth Inference",[22,4114,4115,4116,4119,4120,4123,4124,4127,4128,4131],{},"On Macs, install MLX-LM via ",[26,4117,4118],{},"pip install -U mlx-lm",", then launch OpenAI-compatible server: ",[26,4121,4122],{},"mlx_lm.server jun-song\u002Fsuper-gemma-4-26b-mlx-4bit-v2 --port 8080",", letting it auto-detect the template. Test with ",[26,4125,4126],{},"mlx_lm.generate"," and a prompt at ",[26,4129,4130],{},"--max-tokens 512",". Minimum 24GB unified memory for comfort, 32GB+ preferred; tune wired memory via sysctl if needed. At Q4_K_M quantization, GGUF variant is 16.8GB for broader compatibility.",[22,4133,4134],{},"These steps yield fast, local inference leveraging Gemma's MoE efficiency, enabling seamless tool integration without cloud dependency.",[17,4136,4138],{"id":4137},"pair-with-hermes-or-openclaw-for-terminal-and-assistant-agents","Pair with Hermes or OpenClaw for Terminal and Assistant Agents",[22,4140,4141],{},"Connect MLX\u002FGGUF servers (OpenAI-compatible) to Hermes Agent for terminal-first workflows with tools, memory, MCP, and messaging—select custom OpenAI endpoint, input local URL\u002Fmodel. Hermes leverages Gemma's native function calling for natural agent behavior, not forced adaptations.",[22,4143,4144],{},"For multi-channel assistants, route OpenClaw to the local endpoint as reasoning model, supporting automation and task-running. GGUF works identically via llama.cpp, LM Studio, Jan, or Open WebUI servers.",[22,4146,4147],{},"This stack delivers uncensored, production-like local agents: Gemma base + permissive fine-tune + agent shells, practical for coding\u002Fplanning without refusals.",{"title":72,"searchDepth":73,"depth":73,"links":4149},[4150,4151,4152],{"id":4101,"depth":73,"text":4102},{"id":4111,"depth":73,"text":4112},{"id":4137,"depth":73,"text":4138},[124],{"content_references":4155,"triage":4171},[4156,4157,4159,4161,4162,4165,4167],{"type":3888,"title":3984,"context":88},{"type":3888,"title":4158,"context":3977},"Hermes Agent",{"type":3888,"title":4160,"context":3977},"OpenClaw",{"type":3888,"title":4057,"context":88},{"type":85,"title":4163,"author":4164,"context":88},"SuperGemma-4 26B MLX 4-bit v2","Jun Song",{"type":85,"title":4166,"author":4164,"context":88},"SuperGemma-4 26B GGUF v2 Q4_K_M",{"type":85,"title":4168,"author":4169,"context":4170},"Gemma 4 26B A4B","Google","cited",{"relevance":94,"novelty":3988,"quality":94,"actionability":94,"composite":3989,"reasoning":4172},"Category: AI & LLMs. The article discusses the practical enhancements of SuperGemma-4 for local agent workflows, addressing the audience's need for actionable AI tooling. It provides specific setup instructions and integration tips, making it relevant for developers looking to implement AI features in their products.","\u002Fsummaries\u002F605a7bae59f3f70a-uncensored-supergemma-4-powers-local-agent-workflo-summary","2026-04-16 09:15:03","2026-04-19 03:33:42",{"title":4091,"description":72},{"loc":4173},"605a7bae59f3f70a","https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=VogHvV-M6WE","summaries\u002F605a7bae59f3f70a-uncensored-supergemma-4-powers-local-agent-workflo-summary",[109,108,110,111],"SuperGemma-4 uncensors Gemma 4 26B for text, coding, tool-use, and planning; runs on Apple Silicon via MLX (24GB+ RAM, 46.2 t\u002Fs) or GGUF (16.8GB); integrates with Hermes and OpenClaw for uncensored local agents.",[],"4Sonkeh_FtArw3ake7c1Yv6lKVYLzfA4jtQ8PwyT-C8"]