Por que que aparentemente o genspark degrada a nte...

Question

Por que que aparentemente o genspark degrada a nteligencia quando o request é feito por aqui? mas pelo frontend é normal? está aqui um exemplo: curl 'https://www.google.com/recaptcha/api2/reload?k=6Leq7KYqAAAAAGdd1NaUBJF9dHTPAKP7DcnaRc66'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-type: application/x-protobuffer'
-b '__Secure-3PSID=g.a0000QiBwc8B-cHUyuVOgMRM11FMkyCQOj4D38isOJjXh-0pnA0tvf88G2k6uG-PlKOVVdkyxwACgYKAWgSARASFQHGX2MicklZZgqg36NfEzaym4mWmRoVAUF8yKr1g5Wi18HK_JA7BelfNK8a0076; __Secure-3PAPISID=RW7C0h5t1Pe4I186/AZ5TV4jTz4BRq6_5y; NID=525=nlikg1tZn882MnGYSRghreXoVOXeT1S3mqWQeeiyPHKMtE1sz4LtbzsHCaeIYtOolaqwFsE-VLrdEz-qWlRsL4y7BaW9JfacucKBj9jPfit7E5TaaOJryUIYonu4WSMIpzhPEK0tUypZr7rBMsn_6KeLGxi59jFDVxCRlx0sllakvn4lP_u3RyWSwkqysepjG5byWMUP1XcOv7srlApjpvtDwMor5UgTYCXICJ9HiyZPrZSXkDPuu3mKXnv_gG-WUzyl3_ndFEVVbFfesVT_unDGHg_nmHYBhgImmp4ek-wNgqwdStr5R4yXz594PLxqG-2OuhFJsGa5Bk-wrj77G9ZNgy8P0u_g6ybkINbSiJxHJYHz_dtzkgkzMi7JI_8mELio3l3oMiZyIn_7w1oARD88yjIlvyzTjRGHYdJFc23VytNjAz8thpw4UKMIJtGhPuTmIviOMOf2JoPYebutXl0qWA14nMk-3LuiiBPzkbzaKKu9_sVm0qygMdOhORxKsebsxu_4xA3IqlRO9QOU--VJljtkoo3rYE_jaw2_9Id9Lck_JRNMLjH8t-rCaH7c1YhwbjDeNsI76Q94cYh-t_4aRDWGmXItSfq9I4KVLQAayTbD9bCvNUrEqcrb9AzqVuJQKewNIlCXjsbAJxCRDTP3ncvi1_4ng_UVQJZHsiNFv_n6iiR2BB3SEOSzAFmu9yEvudumAXuXxCE4212S4WbUYiPvfQRROA2u0uuFQc8tS5iXbJaghKqE_Otglits9GIwSyuTJVGatklVu4Vv3BIzrTbkV-jb_aUYinEvci5s7eKcaDyXJdPzR_etSirNDZ-m8nnUgZgmpj6poBD1ntfOqLjs1H3XGQVjOoW22PPDtymglguUlHqhw0AW_BdRwChyiFTKaSydvnzhYg2su2HcR9nG1r7zR_4cNm1QkCCbtyCv5BzF4f3xYl_E5_G6U4Ik1aKw4qGsqI8vr7W07hy8dfyZYBHyNAYzB7Hl5rxq6FsMMIm_gz9aadH6DUgDzVSDnszTwavsAJp0qZ1QEDLEpy5CVvj0Cb4wx5VKtPCxfXvJR5fUSnXVXKat0TmsQBiZmrqSPYYRxnXOrjK0b8t_YuCEHCeJCyhZy_OGkg4q-cezfcSHFdIxtjO7yw7PAQrBftjxXHNCFEnxWhzrVV4YRlID8RejNjnrkbADKsFhZgLY3m3BCgs9TgRy20MV4MMW0RCFBs9MBpQNSmtS_tXnkzRcnMyn-aDh8MXSr6HSh8v7OPRKLmfRonV6wKZ7jt3uFHlxhWaSeu0rP8uWlux46dkMLfwyhjPzQ1vMa__V_1gUAqGs85LkZC-6gPhZ0XazTiWSYp3vu1kxGGXZ_N6wwKYrm01TLToetm6qs8fjyyxevpbN06_5ZH--Q92wHU8dP75yMGBO2i_qM4T5qZ6A3DDnSeFsPeBZUlGb4AwqVHdLHPPaEaa_milRz2UEko8SwYOK5615Gsz7QD7-08hhKvCK_0iPI20mi9jvJHE2cNDOHi6yEMvLym9YRBHVuuahga2_R54NPjbyjkxQcDr9yi4dGd_WDz80yIMa9gtr_9DOknV9Gf60Mghze2UWztdWTdfmadOV6BskLW3NX_rfZ3I4HYPzCEQGWqBdrC8ZQnU; __Secure-3PSIDTS=sidts-CjEB5H03P3YhCM83cF7Ci9jeVNxQ6RaOlbmu8XvPMuVL1XRWm_vwAeEzFVjTLJXEQURLEAA; _Secure-3PSIDCC=AKEyXzWDGwR9P-6DW6hCdVqEUxygqtFUnURvqtrrWFUCUpF2wTpNIH_08dMdrM7wBE09PrM8yQk'
-H 'origin: https://www.google.com'
-H 'priority: u=1, i'
-H 'referer: https://www.google.com/recaptcha/api2/anchor?ar=1&k=6Leq7KYqAAAAAGdd1NaUBJF9dHTPAKP7DcnaRc66&co=aHR0cHM6Ly93d3cuZ2Vuc3BhcmsuYWk6NDQz&hl=pt-BR&v=_mscDd1KHr60EWWbt2I_ULP0&size=invisible&anchor-ms=20000&execute-ms=15000&cb=nxt6mz4awobf'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: cors'
-H 'sec-fetch-site: same-origin'
-H 'sec-fetch-storage-access: active'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0'
--data-raw o5 $g1$ t1755783162 $j22$ l0 $undefined$ '\n(6Leq7KYqAAAAAGdd1NaUBJF9dHTPAKP7DcnaRc66\u0012ù\u000b03AFcWeA4kiUIUYxMIpiVY3xs3nJOzwck752QN66lwU30OILQ66-4HAzTjnP8suuu28e8YXrW6z9-w-IoVza1X4vTLk-zR4QF_OQVXXjC-9kA8PQ-Vs5b7gTwaQe8Qhk-ke3s9E6E-IMy7RcEJLFnF-Q8GXcR3s71_9mDCo0XfzwiBap6aUkMjYIFZ3XvfWixyBzeU6N1-zovDrk4q0R2fbE8ioh0jzK3KGLmb8wbaryYxyf9zjVLrwa0aVAcmx0CyPIuJxtm8f8tIoT_aTGiIVkjQ9nVb7XcnSkdyMct4KkRv2Kx3qrQG5dms6l3AMYDMnLdg6YBGB4l_8KNQocLdsBVj42McYKyCVGHis3l8YrGJ1QB7Uwjijw3iSiMF3aSovh-l8Jp4UrIq6dpHXf3f9BlGUGAENFWAnBZXZ3VTxE9k0hodTh4KMqtNK1T8a3vtRsQqHUBvK1H-YsBdTHa3SFJoeE_GJp8_EnuYuTsnt36WZbNNifw1u7UDMOvj-iPEDwjFTtGqf-20_3o_WsWBfKNkFt7yTKFln0Mr3nhPJTDtDno2Y3voNW28vUtU_RpoIFI1toaq_8eMzNTxxSYxPHI6fJySi_XkwaOygV4-KYckDio1em5acoGZKff5rvU2o5br9qRuj1aSjY706etTvKKXR15oxnsbh-HAz_z2od7e8DPUpB6qZO735Wy5g6FCMnx5aDmvDeiCF1R8P0oLfUUc3XDHLnR6NT6n-euRK1iOjoWO8PaIFAGJDixBNK8npGI30VzpDuw5NqOlzD9FyRdKpAA03Fop3hyT0E7v7G4BvRI5PHKDThyYBl6sXePjO55A2QcxXEfXWF96KZt-L2vD-opPN892AGFTm1TDy1B9nbLXLQGPbUpnWOC3X1L2LodBguPCCdVazfOlGNbqjNF-UeL7Wx-npxkD0E5Mk2zVBJCm7_5DtIRUnVacmiITcSPMre4SnOBopgM_5PZU3kyjNRPTaV3YywrTirgJYsItngrtAgaW8xBnN41yTKCymRy9OzGfBkNmwJtQuxKBD4Nr2Qjv3cTWi7xFW_hNJ59SfbWIyDf-7EgRLI4jJsNdz6W5YVAhknaRNuenl_OcKkQLGnvQ-ajIDr-unikNA8Fm32dCGPyUrrrUKL0y9HWP1_tBabd1KqVK6UlnzBEcEb1r-o7sJUgg91meT4MfkpMo4ZpuTK32gEd4tLGxnHjpRBoGHrYyLxCKrBnn6rHU-B-Sf3vueR1w7M9FDJzsrWiPriIu5cVm99BUTqelK__x1kVBmGDycbYculKlmlLnK_iix8wKSeLJBFTpsm9zN2Vhqij5flRjV1ehwDqL2npYSRBH_KUa-OdaVMrgcrovJ-fYRtiGUMjsGqhFcGFFUZOAW_sS3t8VaoLtDS1ZXuwzWQtV5xhWR-ePJ_DfIcXWBG_HF7DCaGQ0P_gIejAZp1ivrFOTF-S1LaleiOZogRSyhSW9ShG0IQ48n5XoWzuqEnluj7YhHVJD4S7fFeJfLlzUx-RsNvj0Zw\u001a\u0018_mscDd1KHr60EWWbt2I_ULP0"a\u0018\u0001*[\n\n\u0008\u0005\u0010Ö\u000b\u0018¹\u0002 \u0001\n\u0009\u0008\u0005\u0010\u0084¨\u0001\u0018ÿ\u0001\n\u0008\u0008\u000b\u0010´°\u0002\u0018\u000b\n\u0008\u0008\u0006\u0010³°\u0002\u0018\u0013\n\u0009\u0008\u0003\u0010§°\u0002\u0018ö\u0001\n\u0009\u0008\u0001\u0010á©\u0001\u0018\u009d\u0002\n\u0008\u0008\u0002\u0010¨°\u0002\u0018+\n\u000e\u0008\u0004\u0010Õ°\u0002\u0018¾\u0001*\u0003\u0010\u0094^8\u0002' ;
curl 'https://www.genspark.ai/api/user/project_bookmark?project_id=05637783-0274-4021-b1b4-970ddbfcce5e'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-b 'i18n_set=pt-BR; _gcl_au=1.1.62685151.1755263695; _ga=GA1.1.203885191.1755263695; _fbp=fb.1.1755263695001.392179771727185719; ai_user=jmhDbMDM2CtNJcLJVKd5+P|2025-08-15T13:14:55.764Z; session_id=7710c7ce-93c5-46a6-87b9-f357244261e8:67e4794fc2090448254ddd5f9403b4f3a715c1305ada00b8bf8648b4789a9abb; agree_terms=1; c1=23prrvC53bbtTkczLz7xiY8WmSUbjijYnJiMxDM9aQJotIta0Ha%2B8nECeL8T%2BQ%2FCdfePgvv6SBff7FRTQKYDirxgsPa15%2BK2uCsqzuEM%2BaQ1iE8Z; c2=2T8q%2BGSrJGwzDzot; gslogin=1; ai_session=7dZU/OwStpI+Cqq0lLAZ4W|1755779656675|1755783124717; _ga_VL3SBJ0Y0F=GS2.1.s1755783124 $o5$ g1 $t1755783162$ j22 $l0$ h0'
-H 'priority: u=1, i'
-H 'request-id: |dd677a9411a94f9982508dd9d578d848.8f3abd71e6b94259'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: cors'
-H 'sec-fetch-site: same-origin'
-H 'traceparent: 00-dd677a9411a94f9982508dd9d578d848-8f3abd71e6b94259-01'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0' ;
curl 'https://mpc-prod-1-1053047382554.us-central1.run.app/events/3f2102e55cd4604524db906704dfd81de1e24db65a33c67ed8247c8e1e16bff6'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-type: text/plain;charset=UTF-8'
-b 'cee=G%2F9L7cTSQOx4w3ZY%2FwuKkxhiqlH2ynVaR6AUF%2Brz%2FJM%3D.%7B%7D'
-H 'origin: https://www.genspark.ai'
-H 'priority: u=1, i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: cors'
-H 'sec-fetch-site: cross-site'
-H 'sec-fetch-storage-access: active'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0'
--data-raw '{"event_name":"PageView","fb.dynamic_product_ads":{},"custom_data":{},"event_id":"ob3_plugin-set_b2d59a6a63bce91b4f7db1506508fc95509f1da5d8d4fde9ed4c24023690a3db","fb.pixel_id":"984530900272144","fb.advanced_matching":{},"website_context":{"location":"https://www.genspark.ai/agents?id=05637783-0274-4021-b1b4-970ddbfcce5e","referrer":"","isInIFrame":false},"fb.fbp":"fb.1.1755263695001.392179771727185719","event_meta_info":{"experiment_detail":{"name":"use_always_retry","is_exposed":true,"is_in_control":true,"is_in_treatment":false}}}' ;
curl 'https://www.facebook.com/tr/?id=984530900272144&ev=PageView&dl=https%3A%2F%2Fwww.genspark.ai%2Fagents%3Fid%3D05637783-0274-4021-b1b4-970ddbfcce5e&rl=&if=false&ts=1755783164085&sw=1920&sh=1080&v=2.9.225&r=stable&ec=1&o=4126&fbp=fb.1.1755263695001.392179771727185719&eid=ob3_plugin-set_b2d59a6a63bce91b4f7db1506508fc95509f1da5d8d4fde9ed4c24023690a3db&ler=empty&pmd $title$ =Genspark%20-%20Chat%20com%20IA&plt=1244.800000000745&it=1755783124773&coo=false&exp=u3&expv2%5B%5D=pl0&expv2%5B%5D=el0&expv2%5B%5D=bc0&rqm=GET'
-H 'accept: image/avif,image/webp,image/apng,image/svg+xml,image/,/;q=0.8'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'priority: i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: image'
-H 'sec-fetch-mode: no-cors'
-H 'sec-fetch-site: cross-site'
-H 'sec-fetch-storage-access: none'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0' ;
curl 'https://www.facebook.com/privacy_sandbox/pixel/register/trigger/?id=984530900272144&ev=PageView&dl=https%3A%2F%2Fwww.genspark.ai%2Fagents%3Fid%3D05637783-0274-4021-b1b4-970ddbfcce5e&rl=&if=false&ts=1755783164085&sw=1920&sh=1080&v=2.9.225&r=stable&ec=1&o=4126&fbp=fb.1.1755263695001.392179771727185719&eid=ob3_plugin-set_b2d59a6a63bce91b4f7db1506508fc95509f1da5d8d4fde9ed4c24023690a3db&ler=empty&pmd $title$ =Genspark%20-%20Chat%20com%20IA&plt=1244.800000000745&it=1755783124773&coo=false&exp=u3&expv2%5B%5D=pl0&expv2%5B%5D=el0&expv2%5B%5D=bc0&rqm=FGET'
-H 'accept: image/avif,image/webp,image/apng,image/svg+xml,image/,/;q=0.8'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'attribution-reporting-eligible: trigger, event-source;navigation-source'
-H 'attribution-reporting-support: web'
-H 'priority: i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: image'
-H 'sec-fetch-mode: no-cors'
-H 'sec-fetch-site: cross-site'
-H 'sec-fetch-storage-access: none'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0' ;
curl 'https://www.genspark.ai/api/history/log'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-type: application/json'
-b 'i18n_set=pt-BR; _gcl_au=1.1.62685151.1755263695; _ga=GA1.1.203885191.1755263695; _fbp=fb.1.1755263695001.392179771727185719; ai_user=jmhDbMDM2CtNJcLJVKd5+P|2025-08-15T13:14:55.764Z; session_id=7710c7ce-93c5-46a6-87b9-f357244261e8:67e4794fc2090448254ddd5f9403b4f3a715c1305ada00b8bf8648b4789a9abb; agree_terms=1; c1=23prrvC53bbtTkczLz7xiY8WmSUbjijYnJiMxDM9aQJotIta0Ha%2B8nECeL8T%2BQ%2FCdfePgvv6SBff7FRTQKYDirxgsPa15%2BK2uCsqzuEM%2BaQ1iE8Z; c2=2T8q%2BGSrJGwzDzot; gslogin=1; ai_session=7dZU/OwStpI+Cqq0lLAZ4W|1755779656675|1755783124717; _ga_VL3SBJ0Y0F=GS2.1.s1755783124 $o5$ g1 $t1755783162$ j22 $l0$ h0'
-H 'origin: https://www.genspark.ai'
-H 'priority: u=1, i'
-H 'request-id: |dd677a9411a94f9982508dd9d578d848.4dec78ce29ce45e1'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: cors'
-H 'sec-fetch-site: same-origin'
-H 'traceparent: 00-dd677a9411a94f9982508dd9d578d848-4dec78ce29ce45e1-01'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0'
--data-raw '{"type":"agent","agent_type":"moa_chat","key":"05637783-0274-4021-b1b4-970ddbfcce5e","raw_url":"https://www.genspark.ai/agents?id=05637783-0274-4021-b1b4-970ddbfcce5e","favicon_url":"https://cdn1.genspark.ai/user-upload-image/manual/ai_chat.svg","title":"Quatro cidades A, B, C e D estão ligadas por seis pontes distintas da seguinte maneira: – uma ponte "}' ;
curl 'https://westus2-2.in.applicationinsights.azure.com/v2/track'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-type: application/json'
-H 'origin: https://www.genspark.ai'
-H 'priority: u=1, i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: cors'
-H 'sec-fetch-site: cross-site'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0'
--data-raw '[{"time":"2025-08-21T13:32:42.624Z","iKey":"5c09cc03-8ac4-449e-a543-7415b6801d7f","name":"Microsoft.ApplicationInsights.5c09cc038ac4449ea5437415b6801d7f.RemoteDependency","tags":{"ai.user.id":"jmhDbMDM2CtNJcLJVKd5+P","ai.session.id":"7dZU/OwStpI+Cqq0lLAZ4W","ai.device.id":"browser","ai.device.type":"Browser","ai.operation.name":"/agents","ai.operation.id":"78438fc461644bdd86fce4a775537eb7","ai.internal.sdkVersion":"javascript:3.3.3"},"data":{"baseType":"RemoteDependencyData","baseData":{"id":"|78438fc461644bdd86fce4a775537eb7.51670c8968204c76.","ver":2,"name":"POST https://www.google.com/recaptcha/api2/clr?k=6Leq7KYqAAAAAGdd1NaUBJF9dHTPAKP7DcnaRc66","resultCode":"0","duration":"00:00:00.172","success":false,"data":"POST https://www.google.com/recaptcha/api2/clr?k=6Leq7KYqAAAAAGdd1NaUBJF9dHTPAKP7DcnaRc66","target":"www.google.com","type":"Fetch","properties":{"HttpMethod":"POST"},"measurements":{}}}},{"time":"2025-08-21T13:32:42.600Z","iKey":"5c09cc03-8ac4-449e-a543-7415b6801d7f","name":"Microsoft.ApplicationInsights.5c09cc038ac4449ea5437415b6801d7f.RemoteDependency","tags":{"ai.user.id":"jmhDbMDM2CtNJcLJVKd5+P","ai.session.id":"7dZU/OwStpI+Cqq0lLAZ4W","ai.device.id":"browser","ai.device.type":"Browser","ai.operation.name":"/agents","ai.operation.id":"78438fc461644bdd86fce4a775537eb7","ai.internal.sdkVersion":"javascript:3.3.3"},"data":{"baseType":"RemoteDependencyData","baseData":{"id":"|78438fc461644bdd86fce4a775537eb7.42e05dcdca734a3a.","ver":2,"name":"POST /api/copilot/ask","resultCode":"200","duration":"00:00:01.484","success":true,"data":"POST /api/copilot/ask","target":"www.genspark.ai","type":"Fetch","properties":{"HttpMethod":"POST"},"measurements":{}}}},{"time":"2025-08-21T13:32:44.089Z","iKey":"5c09cc03-8ac4-449e-a543-7415b6801d7f","name":"Microsoft.ApplicationInsights.5c09cc038ac4449ea5437415b6801d7f.RemoteDependency","tags":{"ai.user.id":"jmhDbMDM2CtNJcLJVKd5+P","ai.session.id":"7dZU/OwStpI+Cqq0lLAZ4W","ai.device.id":"browser","ai.device.type":"Browser","ai.operation.name":"/agents","ai.operation.id":"dd677a9411a94f9982508dd9d578d848","ai.internal.sdkVersion":"javascript:3.3.3"},"data":{"baseType":"RemoteDependencyData","baseData":{"id":"|dd677a9411a94f9982508dd9d578d848.c51241d187074234.","ver":2,"name":"POST https://mpc-prod-1-1053047382554.us-central1.run.app/events/3f2102e55cd4604524db906704dfd81de1e24db65a33c67ed8247c8e1e16bff6","resultCode":"200","duration":"00:00:00.260","success":true,"data":"POST https://mpc-prod-1-1053047382554.us-central1.run.app/events/3f2102e55cd4604524db906704dfd81de1e24db65a33c67ed8247c8e1e16bff6","target":"mpc-prod-1-1053047382554.us-central1.run.app","type":"Fetch","properties":{"HttpMethod":"POST"},"measurements":{}}}},{"time":"2025-08-21T13:32:44.585Z","iKey":"5c09cc03-8ac4-449e-a543-7415b6801d7f","name":"Microsoft.ApplicationInsights.5c09cc038ac4449ea5437415b6801d7f.Pageview","tags":{"ai.user.id":"jmhDbMDM2CtNJcLJVKd5+P","ai.session.id":"7dZU/OwStpI+Cqq0lLAZ4W","ai.device.id":"browser","ai.device.type":"Browser","ai.operation.name":"/agents","ai.operation.id":"dd677a9411a94f9982508dd9d578d848","ai.internal.sdkVersion":"javascript:3.3.3","ai.internal.snippet":"-"},"data":{"baseType":"PageviewData","baseData":{"ver":2,"name":"Genspark - Quatro cidades A, B, C e D estão ligadas por seis pontes distintas da seguinte maneira: – uma ponte","url":"https://www.genspark.ai/agents?id=05637783-0274-4021-b1b4-970ddbfcce5e","duration":"00:00:00.000","properties":{"refUri":"https://www.genspark.ai/agents?type=moa_chat"},"measurements":{"duration":0},"id":"dd677a9411a94f9982508dd9d578d848"}}}]' ;
curl 'https://westus2-2.in.applicationinsights.azure.com/v2/track'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-type: application/json'
-H 'origin: https://www.genspark.ai'
-H 'priority: u=1, i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: cors'
-H 'sec-fetch-site: cross-site'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0'
--data-raw '[{"time":"2025-08-21T13:32:44.587Z","iKey":"5c09cc03-8ac4-449e-a543-7415b6801d7f","name":"Microsoft.ApplicationInsights.5c09cc038ac4449ea5437415b6801d7f.Metric","tags":{"ai.user.id":"jmhDbMDM2CtNJcLJVKd5+P","ai.session.id":"7dZU/OwStpI+Cqq0lLAZ4W","ai.device.id":"browser","ai.device.type":"Browser","ai.operation.name":"/agents","ai.operation.id":"dd677a9411a94f9982508dd9d578d848","ai.internal.sdkVersion":"javascript:3.3.3"},"data":{"baseType":"MetricData","baseData":{"ver":2,"metrics":[{"name":"PageVisitTime","kind":0,"value":39869,"count":1,"min":39869,"max":39869}],"properties":{"PageName":"Genspark Agents","PageUrl":"https://www.genspark.ai/agents?type=moa_chat"}}}}]' ;
curl 'https://analytics.google.com/g/collect?v=2&tid=G-VL3SBJ0Y0F&gtm=45je58j1v9190289804za200zd9190289804&_p=1755783124379&gcd=13l3l3l3l1l1&npa=0&dma=0&tag_exp=101509157~~103116026~~103200004~~103233427~~104527907~~104528500~~104684208~~104684211~~104948813&cid=203885191.1755263695&ul=pt-br&sr=1920x1080&uaa=x86&uab=64&uafvl=Not%253BA%253DBrand%3B99.0.0.0%7CMicrosoft%2520Edge%3B139.0.3405.102%7CChromium%3B139.0.7258.128&uamb=0&uam=&uap=Windows&uapv=19.0.0&uaw=0&are=1&frm=0&pscdl=noapi&_eu=AAAAAAQ&_s=3&sid=1755783124&sct=5&seg=1&dl=https%3A%2F%2Fwww.genspark.ai%2Fagents%3Ftype%3Dmoa_chat&dt=Genspark%20-%20Chat%20com%20IA&en=copilot_ask&_ee=1&ep.user_s_input=Quatro%20cidades%20A%2C%20B%2C%20C%20e%20D%20est%C3%A3o%20ligadas%20por%20seis%20pontes%20distintas%20da%20seguinte%20maneira%3A%20%E2%80%93%20uma%20ponte%20liga%20A%20e%20B%3B%20%E2%80%93%20uma%20ponte%20liga%20A%20e%20C%3B%20%E2%80%93%20uma%20ponte%20liga%20B%20e%20C%3B%20%E2%80%93%20duas%20pontes%20ligam%20B%20e%20D%3B%20%E2%80%93%20uma%20ponte%20liga%20C%20e%20D.%20Quantos%20caminhos%20s%C3%A3o%20poss%C3%ADveis%20ligando%20todas%20as%20cidades%20e%20passando%20por%20todas%20as%20pontes%20uma%20%C3%BAnica%20vez%2C%20sabendo%20que%20%C3%A9%20permitido%20passar%20em%20uma%20mesma%20cidade%20mais%20de%20uma%20vez%3F&ep.model=gpt-5-high&_et=26706&tfd=41690'
-X 'POST'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-length: 0'
-b '__Secure-3PSID=g.a0000QiBwc8B-cHUyuVOgMRM11FMkyCQOj4D38isOJjXh-0pnA0tvf88G2k6uG-PlKOVVdkyxwACgYKAWgSARASFQHGX2MicklZZgqg36NfEzaym4mWmRoVAUF8yKr1g5Wi18HK_JA7BelfNK8a0076; __Secure-3PAPISID=RW7C0h5t1Pe4I186/AZ5TV4jTz4BRq6_5y; NID=525=nlikg1tZn882MnGYSRghreXoVOXeT1S3mqWQeeiyPHKMtE1sz4LtbzsHCaeIYtOolaqwFsE-VLrdEz-qWlRsL4y7BaW9JfacucKBj9jPfit7E5TaaOJryUIYonu4WSMIpzhPEK0tUypZr7rBMsn_6KeLGxi59jFDVxCRlx0sllakvn4lP_u3RyWSwkqysepjG5byWMUP1XcOv7srlApjpvtDwMor5UgTYCXICJ9HiyZPrZSXkDPuu3mKXnv_gG-WUzyl3_ndFEVVbFfesVT_unDGHg_nmHYBhgImmp4ek-wNgqwdStr5R4yXz594PLxqG-2OuhFJsGa5Bk-wrj77G9ZNgy8P0u_g6ybkINbSiJxHJYHz_dtzkgkzMi7JI_8mELio3l3oMiZyIn_7w1oARD88yjIlvyzTjRGHYdJFc23VytNjAz8thpw4UKMIJtGhPuTmIviOMOf2JoPYebutXl0qWA14nMk-3LuiiBPzkbzaKKu9_sVm0qygMdOhORxKsebsxu_4xA3IqlRO9QOU--VJljtkoo3rYE_jaw2_9Id9Lck_JRNMLjH8t-rCaH7c1YhwbjDeNsI76Q94cYh-t_4aRDWGmXItSfq9I4KVLQAayTbD9bCvNUrEqcrb9AzqVuJQKewNIlCXjsbAJxCRDTP3ncvi1_4ng_UVQJZHsiNFv_n6iiR2BB3SEOSzAFmu9yEvudumAXuXxCE4212S4WbUYiPvfQRROA2u0uuFQc8tS5iXbJaghKqE_Otglits9GIwSyuTJVGatklVu4Vv3BIzrTbkV-jb_aUYinEvci5s7eKcaDyXJdPzR_etSirNDZ-m8nnUgZgmpj6poBD1ntfOqLjs1H3XGQVjOoW22PPDtymglguUlHqhw0AW_BdRwChyiFTKaSydvnzhYg2su2HcR9nG1r7zR_4cNm1QkCCbtyCv5BzF4f3xYl_E5_G6U4Ik1aKw4qGsqI8vr7W07hy8dfyZYBHyNAYzB7Hl5rxq6FsMMIm_gz9aadH6DUgDzVSDnszTwavsAJp0qZ1QEDLEpy5CVvj0Cb4wx5VKtPCxfXvJR5fUSnXVXKat0TmsQBiZmrqSPYYRxnXOrjK0b8t_YuCEHCeJCyhZy_OGkg4q-cezfcSHFdIxtjO7yw7PAQrBftjxXHNCFEnxWhzrVV4YRlID8RejNjnrkbADKsFhZgLY3m3BCgs9TgRy20MV4MMW0RCFBs9MBpQNSmtS_tXnkzRcnMyn-aDh8MXSr6HSh8v7OPRKLmfRonV6wKZ7jt3uFHlxhWaSeu0rP8uWlux46dkMLfwyhjPzQ1vMa__V_1gUAqGs85LkZC-6gPhZ0XazTiWSYp3vu1kxGGXZ_N6wwKYrm01TLToetm6qs8fjyyxevpbN06_5ZH--Q92wHU8dP75yMGBO2i_qM4T5qZ6A3DDnSeFsPeBZUlGb4AwqVHdLHPPaEaa_milRz2UEko8SwYOK5615Gsz7QD7-08hhKvCK_0iPI20mi9jvJHE2cNDOHi6yEMvLym9YRBHVuuahga2_R54NPjbyjkxQcDr9yi4dGd_WDz80yIMa9gtr_9DOknV9Gf60Mghze2UWztdWTdfmadOV6BskLW3NX_rfZ3I4HYPzCEQGWqBdrC8ZQnU; __Secure-3PSIDTS=sidts-CjEB5H03P3YhCM83cF7Ci9jeVNxQ6RaOlbmu8XvPMuVL1XRWm_vwAeEzFVjTLJXEQURLEAA; __Secure-3PSIDCC=AKEyXzVFfl8UZFLanpwKKi4ZVJJTHw_TaOv76HDoj6ZLIdVCOCtNWe4FlCznsfKTIEkIL-XIpf8'
-H 'origin: https://www.genspark.ai'
-H 'priority: u=1, i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: no-cors'
-H 'sec-fetch-site: cross-site'
-H 'sec-fetch-storage-access: active'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0' ;
curl 'https://analytics.google.com/g/collect?v=2&tid=G-VL3SBJ0Y0F&gtm=45je58j1v9190289804za200zd9190289804&_p=1755783124379&gcd=13l3l3l3l1l1&npa=0&dma=0&tag_exp=101509157~~103116026~~103200004~~103233427~~104527907~~104528500~~104684208~~104684211~~104948813&cid=203885191.1755263695&ul=pt-br&sr=1920x1080&uaa=x86&uab=64&uafvl=Not%253BA%253DBrand%3B99.0.0.0%7CMicrosoft%2520Edge%3B139.0.3405.102%7CChromium%3B139.0.7258.128&uamb=0&uam=&uap=Windows&uapv=19.0.0&uaw=0&are=1&frm=0&pscdl=noapi&_eu=AEAAAAQ&_s=4&dl=https%3A%2F%2Fwww.genspark.ai%2Fagents%3Fid%3D05637783-0274-4021-b1b4-970ddbfcce5e&dr=https%3A%2F%2Fwww.genspark.ai%2Fagents%3Ftype%3Dmoa_chat&sid=1755783124&sct=5&seg=1&dt=Genspark%20-%20Quatro%20cidades%20A%2C%20B%2C%20C%20e%20D%20est%C3%A3o%20ligadas%20por%20seis%20pontes%20distintas%20da%20seguinte%20maneira%3A%20%E2%80%93%20uma%20ponte&en=page_view&_et=2744&tfd=46697'
-X 'POST'
-H 'accept: /'
-H 'accept-language: pt-BR,pt;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6'
-H 'content-length: 0'
-b '__Secure-3PSID=g.a0000QiBwc8B-cHUyuVOgMRM11FMkyCQOj4D38isOJjXh-0pnA0tvf88G2k6uG-PlKOVVdkyxwACgYKAWgSARASFQHGX2MicklZZgqg36NfEzaym4mWmRoVAUF8yKr1g5Wi18HK_JA7BelfNK8a0076; __Secure-3PAPISID=RW7C0h5t1Pe4I186/AZ5TV4jTz4BRq6_5y; NID=525=nlikg1tZn882MnGYSRghreXoVOXeT1S3mqWQeeiyPHKMtE1sz4LtbzsHCaeIYtOolaqwFsE-VLrdEz-qWlRsL4y7BaW9JfacucKBj9jPfit7E5TaaOJryUIYonu4WSMIpzhPEK0tUypZr7rBMsn_6KeLGxi59jFDVxCRlx0sllakvn4lP_u3RyWSwkqysepjG5byWMUP1XcOv7srlApjpvtDwMor5UgTYCXICJ9HiyZPrZSXkDPuu3mKXnv_gG-WUzyl3_ndFEVVbFfesVT_unDGHg_nmHYBhgImmp4ek-wNgqwdStr5R4yXz594PLxqG-2OuhFJsGa5Bk-wrj77G9ZNgy8P0u_g6ybkINbSiJxHJYHz_dtzkgkzMi7JI_8mELio3l3oMiZyIn_7w1oARD88yjIlvyzTjRGHYdJFc23VytNjAz8thpw4UKMIJtGhPuTmIviOMOf2JoPYebutXl0qWA14nMk-3LuiiBPzkbzaKKu9_sVm0qygMdOhORxKsebsxu_4xA3IqlRO9QOU--VJljtkoo3rYE_jaw2_9Id9Lck_JRNMLjH8t-rCaH7c1YhwbjDeNsI76Q94cYh-t_4aRDWGmXItSfq9I4KVLQAayTbD9bCvNUrEqcrb9AzqVuJQKewNIlCXjsbAJxCRDTP3ncvi1_4ng_UVQJZHsiNFv_n6iiR2BB3SEOSzAFmu9yEvudumAXuXxCE4212S4WbUYiPvfQRROA2u0uuFQc8tS5iXbJaghKqE_Otglits9GIwSyuTJVGatklVu4Vv3BIzrTbkV-jb_aUYinEvci5s7eKcaDyXJdPzR_etSirNDZ-m8nnUgZgmpj6poBD1ntfOqLjs1H3XGQVjOoW22PPDtymglguUlHqhw0AW_BdRwChyiFTKaSydvnzhYg2su2HcR9nG1r7zR_4cNm1QkCCbtyCv5BzF4f3xYl_E5_G6U4Ik1aKw4qGsqI8vr7W07hy8dfyZYBHyNAYzB7Hl5rxq6FsMMIm_gz9aadH6DUgDzVSDnszTwavsAJp0qZ1QEDLEpy5CVvj0Cb4wx5VKtPCxfXvJR5fUSnXVXKat0TmsQBiZmrqSPYYRxnXOrjK0b8t_YuCEHCeJCyhZy_OGkg4q-cezfcSHFdIxtjO7yw7PAQrBftjxXHNCFEnxWhzrVV4YRlID8RejNjnrkbADKsFhZgLY3m3BCgs9TgRy20MV4MMW0RCFBs9MBpQNSmtS_tXnkzRcnMyn-aDh8MXSr6HSh8v7OPRKLmfRonV6wKZ7jt3uFHlxhWaSeu0rP8uWlux46dkMLfwyhjPzQ1vMa__V_1gUAqGs85LkZC-6gPhZ0XazTiWSYp3vu1kxGGXZ_N6wwKYrm01TLToetm6qs8fjyyxevpbN06_5ZH--Q92wHU8dP75yMGBO2i_qM4T5qZ6A3DDnSeFsPeBZUlGb4AwqVHdLHPPaEaa_milRz2UEko8SwYOK5615Gsz7QD7-08hhKvCK_0iPI20mi9jvJHE2cNDOHi6yEMvLym9YRBHVuuahga2_R54NPjbyjkxQcDr9yi4dGd_WDz80yIMa9gtr_9DOknV9Gf60Mghze2UWztdWTdfmadOV6BskLW3NX_rfZ3I4HYPzCEQGWqBdrC8ZQnU; __Secure-3PSIDTS=sidts-CjEB5H03P3YhCM83cF7Ci9jeVNxQ6RaOlbmu8XvPMuVL1XRWm_vwAeEzFVjTLJXEQURLEAA; __Secure-3PSIDCC=AKEyXzVFfl8UZFLanpwKKi4ZVJJTHw_TaOv76HDoj6ZLIdVCOCtNWe4FlCznsfKTIEkIL-XIpf8'
-H 'origin: https://www.genspark.ai'
-H 'priority: u=1, i'
-H 'sec-ch-ua: "Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"'
-H 'sec-ch-ua-mobile: ?0'
-H 'sec-ch-ua-platform: "Windows"'
-H 'sec-fetch-dest: empty'
-H 'sec-fetch-mode: no-cors'
-H 'sec-fetch-site: cross-site'
-H 'sec-fetch-storage-access: active'
-H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0' Código completo: # -- coding: utf-8 --

GenSpark AI Chat Automator — Captcha Token Automático + Chat CLI

Recursos:

- Automação de chat com a API da GenSpark AI.

- Obtenção automática do 'g_recaptcha_token' via Playwright.

- Configuração de cookies e modelo de IA.

- Interface de chat via linha de comando (CLI) com streaming de respostas.

- Análise de respostas em formato Server-Sent Events (SSE).

Observação: este script usa Playwright, Selenium e requests. Instale as dependências:

pip install "playwright>=1.41" "requests>=2.30" "selenium>=4.15"

playwright install chromium

import asyncio
import json
import sys
import uuid
from typing import Dict, Any, List, Optional

import requests
from playwright.async_api import async_playwright, Route, TimeoutError as PwTimeout
from selenium import webdriver
from selenium.webdriver.chrome.options import Options
from selenium.webdriver.common.by import By
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
import time
import json

==============================================================================

0) CONFIGURAÇÕES — COOKIES E MODELO

==============================================================================

COLOQUE SEUS COOKIES AQUI.

Você pode obtê-los no painel de Desenvolvedor do seu navegador (F12),

na aba "Rede" (Network), selecionando uma requisição para "genspark.ai".

COOKIES: Dict[str, str] = {
"i18n_set": "pt-BR",
"_gcl_au": "1.1.62685151.1755263695",
"_ga": "GA1.1.203885191.1755263695",
"_fbp": "fb.1.1755263695001.392179771727185719",
"ai_user": "jmhDbMDM2CtNJcLJVKd5+P|2025-08-15T13:14:55.764Z",
"session_id": "7710c7ce-93c5-46a6-87b9-f357244261e8:67e4794fc2090448254ddd5f9403b4f3a715c1305ada00b8bf8648b4789a9abb",
"agree_terms": "1",
"c1": "23prrvC53bbtTkczLz7xiY8WmSUbjijYnJiMxDM9aQJotIta0Ha%2B8nECeL8T%2BQ%2FCdfePgvv6SBff7FRTQKYDirxgsPa15%2BK2uCsqzuEM%2BaQ1iE8Z",
"c2": "2T8q%2BGSrJGwzDzot",
"gslogin": "1",
"ai_session": "wWAFgxi+KmF3F9fWVo2sN2|1755772657326|1755772849339",
# "_ga_VL3SBJ0Y0F": "", # Opcional, pode ser omitido
}

Modelo padrão a ser usado na API.

DEFAULT_MODEL = "gpt-5-high"

Configurações para melhorar qualidade das respostas

MAX_TOKENS = 4096
TEMPERATURE = 0.7
TOP_P = 0.9

Token reCAPTCHA - você pode definir aqui se souber o token

Se deixar vazio, o script tentará obter automaticamente via Playwright

HARDCODED_RECAPTCHA_TOKEN = "0cAFcWeA6wfi2U4nvOUDiiDL1989SEId1eT9S5-yQYYRDpgYuckxCqoVUrs9HUYQJPAiYjMlS0eyA7X8GRswe94CcmYmZ3fhN2Ilqdep_w1Ky5_yzrnrUuuUU4Tw9Oo2oKBknvvYYeJlMuJII-dGiSD1jCBwJL8hz5LZQLaf2M6-ZYfp82AUDAMX-PtPOnzcJOe-GWArSxs4E-_vWEz81tFoQA0KLm4zO4026oSrBHCpoIfed4W1f675SKoMbjB3ZC8QnCZXPmF8GEZakT2WAbvAOO0W5nZthgAW-VErOBVIxp-MipiRzVacTLlyciPUvRy4ESo93pXaat5mp5Q2mtcRp0dTxF72rLYVE47Jnvi9T7UHgyr4Tn8oa7VJWbRu6V_MSIbZZ0lc0Nb8g4NzwKwhmdwLoMnvx0LPKtx99JchMYAnNmALc0yHETAqWdIRJRVqD62pf85HRuCf3dsdApml74bjZT2Hp_A0BKer2kh0R1NKierhcsJXkmm5mBILe9YvrKEJMB2EVuZLTRE6Ek8yHqfGtAlIi1qIVPkBuBB1U0fESsIwmxpoDyRi9iPNQc-df1XY6fn8vICJD-nLwo7a2Vn3hiTb_2_mZccfKYxJRw-S4nbFg5av0cGaXLwxx_wYhTDbIIrmjSavolP1Jt9FWM_OEnfa3iFzRzTTG7L0epV3Sa5sNgvjbHLztF8MjlytKEqVIOjVNrE7TKrK1ihWDN9ekmFEmWuEqaUC1bkq5fXUpRtn5zESoZqHSMF3SxLJWLcQoHYqOgYkQ3mfBb8R6_b6OQfV56G8LBbGjW8e6BRjZ6xykhMpOrjSOWzt3DtcXPvGFT6Vgb9A6QHeEewyirNjSz62hzmiekgzf2bc9RgryCI82FIe61YcimbFYr6GweHcPt_Tpym6FODz3tyvdgD6b6oufCOA1nkAzTv0Y6E7I-UOXRUSkYutIbUPl9XU0G1-05AKAqwsW_PnEA4xkkCx0EU1OVAxhTsuDxq7TqPHbF6rkOi4novnhFWWeylG-oI_5s-JE2JqxkTlNbArp--fjtbev3j_aeyeI6TNUWsOJ3pCTeWHhHQDEbLfsk9Z0qrpzfizGB2DtDliT6hd1Ab6Pjxkzqh72bM33EkfeX1q_0KJv4zmoDws5cTPvh1-BFsScwPxNoDUrUo7jrtJbBGaTZrhdhnxjC9DB-c5t7ZsUs_SSUMV1DdhxqsoKKoo4Vy88r8cOi5BjunOlf11WwIdJXTBYeMf_YsKnVcHwdD7U27Rk359jY9975N30G9EYAOdd7rARBZROBsSktbURphtsNa61r5Kx59ClaxhvSYJnAhWd5tGvqb7HujTi5ZX_tetLPnJsMeR5vApxeV3OI7-bIu1-24iQe97qLfXTlMLVjMrnU5fh91Ye3VTd9AKNcG5aYzo8sgMQbV51giJg3IYMWOvZe_nhu9Y4WkJZysQhW3zRjGif2I0_znvAsQ0idbAHqYyGGkk-DdYZH6o5M-LnC-XDwHsGVgANfi9fJP5kPip-zbWGh-gidLreaS3w698DvLxNUqkw_IDk05ZCKZ4ELxFsRV4pWnfO-4AOeaSzf4f6lxwM"

Histórico de conversa para manter contexto

conversation_history = []

URLs da GenSpark

GENSPARK_CHAT_URL = "https://www.genspark.ai/copilot"
GENSPARK_API_URL = "https://www.genspark.ai/api/copilot/ask"

Headers base para as requisições

BASE_HEADERS = {
"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36 Edg/139.0.0.0",
"Accept": "/",
"Accept-Language": "pt-BR,pt;q=0.9,en;q=0.8,en-US;q=0.7",
"Accept-Encoding": "gzip, deflate, br, zstd",
"Origin": "https://www.genspark.ai",
"Referer": GENSPARK_CHAT_URL,
"Sec-Ch-Ua": '"Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"',
"Sec-Ch-Ua-Mobile": "?0",
"Sec-Ch-Ua-Platform": '"Windows"',
"Sec-Ch-Ua-Platform-Version": '"15.0.0"',
"Sec-Ch-Ua-Arch": '"x86"',
"Sec-Ch-Ua-Bitness": '"64"',
"Sec-Ch-Ua-Full-Version": '"139.0.0.0"',
"Sec-Ch-Ua-Full-Version-List": '"Not;A=Brand";v="99.0.0.0", "Microsoft Edge";v="139.0.0.0", "Chromium";v="139.0.0.0"',
"Sec-Fetch-Dest": "empty",
"Sec-Fetch-Mode": "cors",
"Sec-Fetch-Site": "same-origin",
"Sec-Fetch-User": "?1",
"Sec-Fetch-Storage-Access": "active",
"Content-Type": "application/json",
"Priority": "u=1, i",
"Upgrade-Insecure-Requests": "1",
}

==============================================================================

1) OBTENÇÃO AUTOMÁTICA DO g_recaptcha_token (via Playwright)

==============================================================================

async def get_recaptcha_token(cookies: Dict[str, str], headless: bool = True) -> Optional[str]:
print("[Playwright] Iniciando navegador para obter token reCAPTCHA...")
token_captured = asyncio.Event()
captured_token: Optional[str] = None

text
async def route_handler(route: Route):
    nonlocal captured_token
    req = route.request
    # Intercepta o POST do chat para extrair o token
    if req.url == GENSPARK_API_URL and req.method == "POST":
        try:
            data = await req.post_data_json()
            tok = data.get("g_recaptcha_token")
            if tok:
                print("[Playwright] Token reCAPTCHA capturado com sucesso.")
                captured_token = tok
                token_captured.set()
                await route.abort()  # aborta para não consumirmos quota/ação
                return
        except Exception as e:
            print(f"[Playwright] Erro ao processar POST do chat: {e}")
    await route.continue_()

p = browser = context = page = None
try:
    p = await async_playwright().start()
    browser = await p.chromium.launch(headless=headless, args=["--disable-blink-features=AutomationControlled"])
    context = await browser.new_context(
        user_agent=BASE_HEADERS["User-Agent"],
        locale="pt-BR",  # combina com seus cookies
        timezone_id="America/Sao_Paulo",  # opcional, mas ajuda a parecer "humano"
        viewport={"width": 1366, "height": 768}
    )

    # Aplica cookies (sem depender do placeholder em inglês)
    await context.add_cookies([
        {"name": k, "value": v, "domain": ".genspark.ai", "path": "/"}
        for k, v in cookies.items() if v
    ])

    page = await context.new_page()
    await page.route("**/*", route_handler)

    # Carrega a página e espera a app SPA estabilizar
    await page.goto(GENSPARK_CHAT_URL, wait_until="domcontentloaded")
    await page.wait_for_load_state("networkidle")  # aguarda bundles/SPA

    # Alguns sites mostram banner de cookies/termos; tente fechar se existir (opcional e tolerante a erro)
    try:
        # Procura qualquer botão com texto "Aceitar/Accept/OK" (português/inglês)
        btn = page.get_by_role("button", name=re.compile(r"(aceit|accept|ok|entendi)", re.I)).first
        if await btn.is_visible():
            await btn.click()
    except Exception:
        pass

    # Localizador do campo de entrada — agnóstico de idioma e do tipo (textarea ou contenteditable)
    # Preferimos role="textbox" (Playwright mapeia automaticamente).
    textbox = page.get_by_role("textbox").first

    # Fallback: procura manualmente por elementos comuns de chat
    async def ensure_textbox():
        nonlocal textbox
        try:
            await textbox.wait_for(state="visible", timeout=10000)
            return True
        except PwTimeout:
            pass
        # Tenta padrões alternativos
        candidates = [
            'textarea',
            '[contenteditable="true"]',
            'div[contenteditable="true"]',
            'form textarea',
        ]
        for sel in candidates:
            loc = page.locator(sel).first
            try:
                await loc.wait_for(state="visible", timeout=3000)
                textbox = loc
                return True
            except PwTimeout:
                continue
        return False

    ok = await ensure_textbox()
    if not ok:
        raise PwTimeout("Campo de entrada do chat não encontrado (nem textbox/textarea/contenteditable).")

    # Digita e envia a mensagem para disparar o POST com o token
    try:
        await textbox.click()
        # Algumas implementações de contenteditable não suportam fill()
        try:
            await textbox.fill("Olá")
        except Exception:
            await page.keyboard.type("Olá", delay=50)

        # Tente clicar em um botão "Enviar/Send" se existir; senão, Enter
        try:
            send_btn = page.get_by_role("button", name=re.compile(r"(enviar|send|ask|falar)", re.I)).first
            if await send_btn.is_visible():
                await send_btn.click()
            else:
                await page.keyboard.press("Enter")
        except Exception:
            await page.keyboard.press("Enter")
    except Exception as e:
        raise RuntimeError(f"Falha ao interagir com a UI do chat: {e}")

    # Aguarda o token capturado pela rota
    await asyncio.wait_for(token_captured.wait(), timeout=45)
    return captured_token

except asyncio.TimeoutError:
    print("[Playwright] ERRO: Timeout ao esperar pelo token. A página pode ter mudado, login expirou ou há um desafio adicional (ex.: captcha visível).")
    return None
except Exception as e:
    print(f"[Playwright] ERRO: Ocorreu uma exceção inesperada: {e}")
    return None
finally:
    if context:
        await context.close()
    if browser:
        await browser.close()
    if p:
        await p.stop()
    print("[Playwright] Navegador finalizado.")

==============================================================================

2) CLIENTE DA API GENSPARK

==============================================================================

def build_api_payload(prompt: str, model: str, g_recaptcha_token: str, include_history: bool = True) -> Dict[str, Any]:
"""Constrói o corpo (payload) da requisição para a API de chat."""
global conversation_history

text
# Adicionar a mensagem atual ao histórico
conversation_history.append({"role": "user", "content": prompt})

# Limitar o histórico para evitar payloads muito grandes (últimas 20 mensagens)
if len(conversation_history) > 20:
    conversation_history = conversation_history[-20:]

# Usar todo o histórico ou apenas a mensagem atual
messages = conversation_history if include_history else [{"role": "user", "content": prompt}]

return {
    "type": "COPILOT_MOA_CHAT",
    "current_query_string": "type=COPILOT_MOA_CHAT",
    "project_id": None,
    "messages": messages,
    "user_s_input": prompt,
    "action_params": {},
    "extra_data": {
        "models": [model],
        "run_with_another_model": False,
        "request_web_knowledge": True,
        "speed_mode": False,
        "writingContent": None,
        "temperature": TEMPERATURE,
        "top_p": TOP_P,
        "max_tokens": MAX_TOKENS,
    },
    "g_recaptcha_token": g_recaptcha_token,
}

def process_sse_stream(response: requests.Response) -> str:
"""
Processa a resposta SSE (Server-Sent Events), imprime os pedaços de texto
em tempo real e retorna a resposta completa.
"""
full_text = ""
try:
for line in response.iter_lines(decode_unicode=True):
if line.startswith("data:"):
data_str = line[len("data:"):].strip()
if not data_str or data_str == "[DONE]":
continue
try:
data_json = json.loads(data_str)
# Heurística para encontrar o pedaço de texto
chunk = data_json.get("content", "") or data_json.get("text", "")
if isinstance(chunk, str):
print(chunk, end="", flush=True)
full_text += chunk
except json.JSONDecodeError:
# Ignora linhas que não são JSON válido
pass
except Exception as e:
print(f"\n[API] Erro durante o streaming da resposta: {e}")

text
print() # Nova linha no final
return full_text

def send_chat_request_selenium(prompt: str, cookies: Dict[str, str], model: str, token: str) -> str:
"""Envia a requisição de chat usando Selenium (navegador real) para evitar detecção."""
print("[Selenium] Iniciando navegador para fazer requisição...")

text
driver = None
try:
    # Configurar opções do Chrome
    options = Options()
    options.add_argument("--disable-blink-features=AutomationControlled")
    options.add_experimental_option("excludeSwitches", ["enable-automation"])
    options.add_experimental_option("useAutomationExtension", False)
    options.add_argument("--disable-extensions")
    options.add_argument("--no-sandbox")
    options.add_argument("--disable-dev-shm-usage")
    options.add_argument("--disable-gpu")
    options.add_argument("--window-size=1366,768")
    options.add_argument(f"--user-agent={BASE_HEADERS['User-Agent']}")

    # Inicializar driver
    driver = webdriver.Chrome(options=options)

    # Definir cookies
    driver.get("https://www.genspark.ai")
    for k, v in cookies.items():
        if v:
            driver.add_cookie({"name": k, "value": v, "domain": ".genspark.ai"})

    # Construir payload com histórico
    payload = build_api_payload(prompt, model, token, include_history=True)

    # Debug: mostrar histórico sendo enviado
    print(f"[DEBUG] Enviando histórico com {len(payload['messages'])} mensagens")
    for i, msg in enumerate(payload['messages']):
        print(f"[DEBUG] Msg {i+1}: {msg['role']} - {msg['content'][:30]}...")

    # Executar requisição via JavaScript
    script = f"""
    return fetch('{GENSPARK_API_URL}', {{
        method: 'POST',
        headers: {{
            'Content-Type': 'application/json',
            'User-Agent': '{BASE_HEADERS["User-Agent"]}',
            'Accept': '*/*',
            'Accept-Language': 'pt-BR,pt;q=0.9,en;q=0.8,en-US;q=0.7',
            'Accept-Encoding': 'gzip, deflate, br, zstd',
            'Origin': 'https://www.genspark.ai',
            'Referer': 'https://www.genspark.ai/copilot',
            'Sec-Ch-Ua': '"Not;A=Brand";v="99", "Microsoft Edge";v="139", "Chromium";v="139"',
            'Sec-Ch-Ua-Mobile': '?0',
            'Sec-Ch-Ua-Platform': '"Windows"',
            'Sec-Fetch-Dest': 'empty',
            'Sec-Fetch-Mode': 'cors',
            'Sec-Fetch-Site': 'same-origin'
        }},
        body: JSON.stringify({json.dumps(payload)})
    }}).then(response => response.text());
    """

    # Executar e aguardar resposta
    response_text = driver.execute_script(script)

    # Processar resposta (simular streaming)
    try:
        # Tentar parse do JSON se for uma resposta SSE
        lines = response_text.split('\n')
        full_text = ""
        for line in lines:
            if line.startswith("data:"):
                data_str = line[len("data:"):].strip()
                if not data_str or data_str == "[DONE]":
                    continue
                try:
                    data_json = json.loads(data_str)
                    chunk = data_json.get("content", "") or data_json.get("text", "")
                    if isinstance(chunk, str):
                        print(chunk, end="", flush=True)
                        full_text += chunk
                except json.JSONDecodeError:
                    pass
        print()  # Nova linha

        # Adicionar resposta da IA ao histórico
        if full_text.strip():
            conversation_history.append({"role": "assistant", "content": full_text.strip()})

        return full_text
    except Exception:
        # Se não for SSE, retornar texto bruto
        print(response_text)
        return response_text

except Exception as e:
    print(f"\n[Selenium] ERRO: {e}")
    return ""
finally:
    if driver:
        driver.quit()
        print("[Selenium] Navegador finalizado.")

def get_recaptcha_token_selenium(cookies: Dict[str, str]) -> Optional[str]:
"""Obtém o token reCAPTCHA automaticamente usando Selenium."""
print("[Selenium] Tentando obter token reCAPTCHA automaticamente...")

text
driver = None
try:
    # Configurar opções do Chrome
    options = Options()
    options.add_argument("--disable-blink-features=AutomationControlled")
    options.add_experimental_option("excludeSwitches", ["enable-automation"])
    options.add_experimental_option("useAutomationExtension", False)
    options.add_argument("--disable-extensions")
    options.add_argument("--no-sandbox")
    options.add_argument("--disable-dev-shm-usage")
    options.add_argument("--disable-gpu")
    options.add_argument("--window-size=1366,768")
    options.add_argument(f"--user-agent={BASE_HEADERS['User-Agent']}")

    # Inicializar driver
    driver = webdriver.Chrome(options=options)

    # Definir cookies
    driver.get("https://www.genspark.ai")
    for k, v in cookies.items():
        if v:
            driver.add_cookie({"name": k, "value": v, "domain": ".genspark.ai"})

    # Recarregar a página com os cookies
    driver.get(GENSPARK_CHAT_URL)
    time.sleep(3)  # Aguardar carregamento

    # Tentar encontrar e interagir com o chat
    try:
        # Procurar campo de entrada
        input_selectors = [
            'textarea',
            '[contenteditable="true"]',
            'div[contenteditable="true"]',
            'form textarea',
            'input[type="text"]',
            '.chat-input',
            '.message-input'
        ]

        textbox = None
        for selector in input_selectors:
            try:
                elements = driver.find_elements(By.CSS_SELECTOR, selector)
                for element in elements:
                    if element.is_displayed() and element.is_enabled():
                        textbox = element
                        break
                if textbox:
                    break
            except:
                continue

        if not textbox:
            print("[Selenium] Campo de entrada não encontrado")
            return None

        # Focar no campo
        textbox.click()
        time.sleep(1)

        # Digitar mensagem de teste
        textbox.send_keys("teste")
        time.sleep(1)

        # Procurar botão de enviar
        send_selectors = [
            'button[type="submit"]',
            'button:has-text("enviar")',
            'button:has-text("send")',
            'button:has-text("ask")',
            '.send-button',
            '.submit-button'
        ]

        send_button = None
        for selector in send_selectors:
            try:
                elements = driver.find_elements(By.CSS_SELECTOR, selector)
                for element in elements:
                    if element.is_displayed() and element.is_enabled():
                        send_button = element
                        break
                if send_button:
                    break
            except:
                continue

        if send_button:
            send_button.click()
        else:
            # Tentar Enter
            textbox.send_keys('\n')

        # Aguardar um pouco para a requisição ser enviada
        time.sleep(5)

        # Verificar se conseguimos interceptar o token
        # Como não podemos interceptar requisições facilmente no Selenium,
        # vamos assumir que se chegou aqui, o token foi usado
        # Para uma implementação completa, seria necessário usar um proxy
        print("[Selenium] Não foi possível interceptar o token automaticamente")
        return None

    except Exception as e:
        print(f"[Selenium] Erro ao interagir com a página: {e}")
        return None

except Exception as e:
    print(f"[Selenium] Erro geral: {e}")
    return None
finally:
    if driver:
        driver.quit()

def send_chat_request(prompt: str, cookies: Dict[str, str], model: str, token: str) -> str:
"""Envia a requisição de chat usando Selenium (fallback para requests)."""
return send_chat_request_selenium(prompt, cookies, model, token)

==============================================================================

3) CLI — CONVERSA INTERATIVA

==============================================================================

async def chat_cli():
"""Loop principal para a interface de linha de comando."""
print("=== GenSpark AI CLI ===")
print("Comandos: /quit, /exit, /clear, /history")
print(f"Modelo atual: {DEFAULT_MODEL}")

text
while True:
    try:
        user_input = input("\nVocê> ").strip()
    except (KeyboardInterrupt, EOFError):
        print("\nEncerrando.")
        break

    if not user_input:
        continue
    
    if user_input.lower() in ("/quit", "/exit"):
        print("Até mais!")
        break
    elif user_input.lower() == "/clear":
        conversation_history = []
        print("Histórico de conversa limpo!")
        continue
    elif user_input.lower() == "/history":
        if not conversation_history:
            print("Histórico vazio.")
        else:
            print(f"Histórico ({len(conversation_history)} mensagens):")
            for i, msg in enumerate(conversation_history, 1):
                role = "Você" if msg["role"] == "user" else "IA"
                print(f"{i}. {role}: {msg['content'][:50]}...")
        continue
    
    # 1. Obter o token reCAPTCHA para esta mensagem
    # Primeiro tentar obter automaticamente com Selenium
    token = None
    try:
        token = get_recaptcha_token_selenium(COOKIES)
    except Exception as e:
        print(f"[INFO] Não foi possível obter token automaticamente: {e}")

    # Se não conseguiu automaticamente, usar hardcoded se disponível
    if not token and HARDCODED_RECAPTCHA_TOKEN:
        token = HARDCODED_RECAPTCHA_TOKEN
        print("[INFO] Usando token hardcoded")

    # Se ainda não tem token, tentar com Playwright
    if not token:
        try:
            token = await get_recaptcha_token(COOKIES)
        except Exception as e:
            print(f"[INFO] Não foi possível obter token com Playwright: {e}")

    if not token:
        print("[ERRO] Não foi possível obter o token reCAPTCHA. Tente novamente.")
        print("[DICA] Defina HARDCODED_RECAPTCHA_TOKEN no início do script ou verifique se os cookies estão corretos e válidos.")
        continue

    # 2. Enviar a mensagem para a API
    print(f"GenSpark ({DEFAULT_MODEL})> ", end="", flush=True)
    send_chat_request(user_input, COOKIES, DEFAULT_MODEL, token)

==============================================================================

4) PONTO DE ENTRADA (MAIN)

==============================================================================

if name == "main":
if not COOKIES.get("session_id") or not COOKIES.get("ai_user"):
print("ERRO DE CONFIGURAÇÃO:")
print("Por favor, preencha os valores de 'session_id' e 'ai_user' no dicionário COOKIES no início do script.")
sys.exit(1)

text
try:
    asyncio.run(chat_cli())
except Exception as e:
    print(f"\nOcorreu um erro fatal no loop principal: {e}")
    # Tenta executar no loop de evento existente, caso esteja em um ambiente como IPython/Jupyter
    try:
        loop = asyncio.get_running_loop()
        if loop.is_running():
            print("Tentando agendar em um loop de evento já em execução...")
            task = loop.create_task(chat_cli())
            loop.run_until_complete(task)
    except RuntimeError:
        pass # Se não houver loop, o programa termina

Answer 1

Pelo que você mostrou, o “pior desempenho” não é o modelo “ficando burro”, e sim mudança de rota/recursos no back-end quando a chamada não vem do fluxo oficial do front. Em plataformas como a GenSpark, o servidor classifica a confiança do cliente (origem, sessão, recaptcha, feature-flags, etc.). Se algo não bate, ele pode:

Trocar o roteamento do modelo
- O front costuma usar um router (p.ex. combinações como “gpt-5-high” + fallback ou run_with_another_model=true), prompt de sistema oculto e ferramentas internas.
- Na sua chamada você fixa "models":["gpt-5-high"] e "run_with_another_model": false. Se a UI normalmente habilita o router, você perde upgrades automáticos, contextos maiores ou post-processing.
Desligar ferramentas/contexto que a UI injeta
- O front normalmente envia metadados, prompts de sistema e tool grants invisíveis (RAG interno, web-knowledge, memórias da sessão, etc.).
- No seu primeiro curl, veio "request_web_knowledge": false. Se no site isso vai “true” ou existe RAG interno, a qualidade cai mesmo com o mesmo modelo.
Cair em trilha de “low-trust” por reCAPTCHA/origem
- Tokens do reCAPTCHA v2/v3 são vinculados ao domínio/ação e expiram em minutos. Reutilizar/“hardcode” token, ou obter token em uma ação diferente, pode passar mas classifica a requisição como baixa confiança, limitando ferramentas ou roteando para um caminho mais restritivo.
- Cabe notar: não posso ajudar a contornar reCAPTCHA ou automatizar sua resolução. Além de anti-abuso, isso geralmente fere os Termos.
Headers/sessão diferentes da UI
- A UI costuma mandar cabeçalhos, cookies e feature flags (experimentos A/B) que liberam capacidades. Sem eles, você pode ficar fora de experimentos “pro” (ex.: use_always_retry, long context, etc.).
- Chamar via fetch injetado pelo Selenium nem sempre replica os mesmos CSRF/XSRF, Referer, Sec-CH-UA*, e principalmente o contexto de projeto (se a UI envia project_id/workspace).
Parâmetros de geração “apertados”
- Você fixa temperature, top_p e max_tokens; o front pode ajustar dinamicamente por tipo de pergunta. Para problemas combinatórios/algébricos isso pesa.

Como diagnosticar e corrigir (sem burlar proteções)

Capture e compare um request 100% real do front vs o seu:
- body: liste todos os campos extras (router, request_web_knowledge, project_id, action_params, system hints).
- headers: veja se existe algum x-*/CSRF, Referer, Origin, e Client Hints que a sua automação não replica.
Deixe o router decidir o modelo se a plataforma permite:
- Use o mesmo payload do front para extra_data (p.ex. habilite run_with_another_model e não fixe demais o modelo).
Não reutilize/forge tokens reCAPTCHA. Use os métodos suportados (SDK ou API oficial server-to-server) para evitar cair na trilha “low-trust”.
Habilite as mesmas features que a UI liga:
- Se o front manda request_web_knowledge: true (ou outra flag RAG), replique.
- Se há project_id/workspace, mande igual — isso muda fontes e system prompt.
Verifique limites de contexto e max_tokens: alinhe com a UI.
Evite Selenium/Playwright para a chamada final. Autentique e chame via API oficial (com credenciais próprias), em vez de depender de cookies da sessão web.

Checklist rápido

Usar endpoint/SDK oficial e credencial própria (nada de cookies da web).
Copiar payload exato da UI (incluindo request_web_knowledge, project_id, action_params, router).
Remover HARDCODED_RECAPTCHA_TOKEN e qualquer automação de captcha.
Igualar headers críticos (Origin/Referer, CSRF) se a API exigir – mas o ideal é server-to-server.
Não fixar run_with_another_model=false—deixe o roteador trabalhar.
Ajustar max_tokens/temperature para o tipo de tarefa.

Se você quiser, eu comparo dois payloads (um capturado da UI e o seu) e te devolvo exatamente o que está faltando/extra — basta colar o JSON “limpo” (sem cookies/tokens/sigilos).

Frage

GenSpark AI Chat Automator — Captcha Token Automático + Chat CLI

Recursos:

- Automação de chat com a API da GenSpark AI.

- Obtenção automática do 'g_recaptcha_token' via Playwright.

- Configuração de cookies e modelo de IA.

- Interface de chat via linha de comando (CLI) com streaming de respostas.

- Análise de respostas em formato Server-Sent Events (SSE).

Observação: este script usa Playwright, Selenium e requests. Instale as dependências:

pip install "playwright>=1.41" "requests>=2.30" "selenium>=4.15"

playwright install chromium

==============================================================================

0) CONFIGURAÇÕES — COOKIES E MODELO

==============================================================================

COLOQUE SEUS COOKIES AQUI.

Você pode obtê-los no painel de Desenvolvedor do seu navegador (F12),

na aba "Rede" (Network), selecionando uma requisição para "genspark.ai".

Modelo padrão a ser usado na API.

Configurações para melhorar qualidade das respostas

Token reCAPTCHA - você pode definir aqui se souber o token

Se deixar vazio, o script tentará obter automaticamente via Playwright

Histórico de conversa para manter contexto

URLs da GenSpark

Headers base para as requisições

==============================================================================

1) OBTENÇÃO AUTOMÁTICA DO g_recaptcha_token (via Playwright)

==============================================================================

==============================================================================

2) CLIENTE DA API GENSPARK

==============================================================================

==============================================================================

3) CLI — CONVERSA INTERATIVA

==============================================================================

==============================================================================

4) PONTO DE ENTRADA (MAIN)

==============================================================================

Gedanke

Antwort

Como diagnosticar e corrigir (sem burlar proteções)

Checklist rápido

Diesen Q&A teilen