I-Supercharge elawulwa yimithwalo yomsebenzi nge-Claude Code kanye ne-Amazon Bedrock

Ukukhishwa kwe-Anthropic Claude Sonnet 4.5 Esifundeni se-AWS GovCloud (US) kwethula i-on-ramp eqondile yokuthuthukiswa okusizwa yi-AI ngemithwalo yemisebenzi enezidingo zokuthobela imithetho.

Kulokhu okuthunyelwe, sihlola ukuthi singamhlanganisa kanjani uClaude Sonnet 4.5 ku-Amazon Bedrock ku-AWS GovCloud (US) ne-Claude Code, umsizi wokubhala amakhodi we-ejenti okhishwe yi-Anthropic. Lesi sixazululo sinika amandla ukuhamba komsebenzi okusizwa yi-AI okusiza ukuthuthukisa imisebenzi yentuthuko yansuku zonke kuyilapho kusiza izinhlangano ukuthi zigcine ukuhambisana okuqinile nokuhlangabezana nezidingo zokulawula ezifana ne-International Traffic in Arms Regulations (ITAR), Impact Level (IL) 4 noma 5, kanye nokuthobelana okuphezulu kwe-FedRAMP.

I-Amazon Bedrock ku-AWS GovCloud (US)

Izifunda ze-AWS GovCloud (US) ziklanyelwe ngqo amakhasimende ase-US anomthwalo wemisebenzi odinga ukugunyazwa okuqinile kokuthobela imithetho. Lokhu kuvumela imithwalo yomsebenzi ye-AI ekhiqizayo ukuthi isetshenziswe ezindaweni ezibucayi kuyilapho kugcinwa izilawuli zokuthobela ezifanele. I-Amazon Bedrock yakhiwe ngezici ezigxile kwezokuphepha, okuhlanganisa:

• Ukuvikelwa kwedatha eyakhelwe ngaphakathi lapho okuqukethwe kwekhasimende kungagciniwe, kungalogwa, noma kungasetshenziswa ukuqeqesha amamodeli e-AWS noma kwabelwane ngawo nezinkampani zangaphandle.
• I-FedRAMP High kanye ne-DoD Cloud Computing (CC) SRG Impact Level (IL) 4/5 izindlela zokugunyazwa, ezisekela izidingo zokuthobela ze-ejensi kahulumeni.
• Ukuhlanganiswa nezilawuli zokuphepha ezikhona kanye nezinhlaka zokuthobela ezitholakala ku-AWS GovCloud (US), ukugcina izindinganiso zokuphepha eziphezulu njengezinye Izifunda ze-AWS kuyilapho kunikezwa izindlela zokugunyazwa ezengeziwe.

Claude Ikhodi

I-Claude Code ingumsizi wokubhala amakhodi we-Anthropic we-AI enikwa amandla amamodeli akamuva esisekelo, njenge-Anthropic Claude Sonnet 4.5, etholakala ku-Amazon Bedrock e-AWS GovCloud (US-West) kanye ne-GovCloud (US-East) nge-US-GOV cross-Region Inference. Isebenza ngokuqondile kutheminali yakho, indawo oyintandokazi yakho yokuthuthukisa edidiyelwe (ama-IDE) njenge-VS Code ne-JetBrains, futhi ngemuva ine-Claude Agent SDK. Ikhodi ye-Claude ingaqonda umongo wephrojekthi yakho futhi ithathe izinyathelo ezingaphakathi ngaphandle kokudinga ukuthi wengeze amazwibela ekhodi akhiqiziwe kuphrojekthi. Ngokungafani nabasizi bendabuko bokubhala amakhodi, i-Claude Code ingakwazi:

• Bhala ikhodi futhi ulungise iziphazamisi ezihlanganisa amafayela amaningi ku-codebase yakho
• Phendula imibuzo mayelana nesakhiwo sekhodi yakho kanye nengqondo
• Yenza futhi ulungise izivivinyo, ukuhlanganisa, neminye imiyalo
• Sesha umlando we-Git, xazulula ukuhlanganisa izingxabano, futhi udale ukuzibophezela nokudonsa izicelo
• Nweba kwamanye amathuluzi omugqa womyalo, njenge-AWS Command Line Interface (AWS CLI), Terraform, K8s, kanye namaseva e-MCP

Ukuze ufunde kabanzi, bheka izindatshana ze-Anthropic: Okokufundisa kwekhodi ka-Claude kanye nekhodi ka-Claude: Izindlela ezihamba phambili zokubhala amakhodi kwe-ejenti.

Uhlolojikelele lwesixazululo: Zama i-Claude Code nge-Amazon Bedrock ku-GovCloud

Lesi sigaba sinikeza imiyalelo yesinyathelo ngesinyathelo yokusetha nokulungisa Ikhodi ye-Claude ukuze isebenze ne-Amazon Bedrock ku-AWS GovCloud, okuhlanganisa izimfuneko, imiyalo yokufaka, ukumisa imvelo, nezinyathelo zokuqinisekisa.

Okudingekayo

Ngaphambi kokuthi uqale, qiniseka ukuthi unokulandelayo endaweni:

Ukwandisa ukusebenza kahle kokuqondisisa, ukugcinwa kwesikhashana okusheshayo kuvulwa ngokuzenzakalelayo kumamodeli asekelwayo kanye Nezifunda ze-AWS—kuhlanganise noClaude Sonnet 4.5.

Setha Ikhodi kaClaude noClaude Sonnet 4.5 ku-Amazon Bedrock

Ngemva kokumisa i-AWS CLI ngemininingwane yakho, kulesi sibonelo ufaka i-Claude Code ye-macOS, Linux, kanye ne-WSL (izindlela zokufaka ezengeziwe zingatholakala ku-Claude Code Amadokhumenti):

1. Faka Ikhodi ye-Claude:

macOS, Linux, WSL:
curl -fsSL  | bash

2. Lungiselela i-Amazon Bedrock:

export AWS_REGION=’us-gov-west-1’
export CLAUDE_CODE_USE_BEDROCK=1
export ANTHROPIC_MODEL=’anthropic.claude-sonnet-4-5-20250929-v1:0’

3. Zulazulela kuphrojekthi yakho:

4. Yethula Ikhodi kaClaude

5. Qinisekisa ukuthi i-Claude Code iyasebenza ngokubheka i- Welcome to Claude Code! umyalezo kutheminali yakho.

Ukuze uqinisekise imodeli yakho eshiwo, sebenzisa ifayela le- /status umyalo. Lokhu kuzobonisa imodeli ye-ID echazwe esinyathelweni sesi-2. Manje usuxhume ngempumelelo Ikhodi ye-Claude ku-Amazon Bedrock ku-GovCloud.

Ukuze ufunde kabanzi mayelana nendlela yokumisa Ikhodi ye-Claude ye-Amazon Bedrock, bheka Ikhodi ye-Claude ku-Amazon Bedrock.

Okucatshangelwayo lapho uthumela i-Claude Code enhlanganweni yakho

Njengoba i-Claude Code manje itholakala ngokuvamile, amakhasimende amaningi acabangela izinketho zokuthumela ku-AWS ukuze asebenzise amandla ayo okubhala athuthukisiwe. Ukuze uthole ukuthunyelwa, cabangela isakhiwo sakho esiyisisekelo sokuphepha, ukubusa nokuhambisana:

Sebenzisa Isikhungo Somazisi se-AWS IAM ukuphatha ubuwena kanye nokufinyelela ku-Claude Code. Lokhu kuqinisekisa ukuthi onjiniyela abagunyaziwe kuphela abanokufinyelela. Ukwengeza, besebenzisa i-IAM Identity Center, abathuthukisi bangakwazi ukufinyelela izinsiza ezinobufakazi besikhashana, obusekelwe endimeni, okunciphisa isidingo sokhiye bokufinyelela abangashintshi futhi basize ekuthuthukiseni ukuvikeleka. Ngaphambi kokuvula Ikhodi ye-Claude, qiniseka ukuthi ulungiselela i-AWS CLI ukusebenzisa iphrofayela ye-IAM Identity Center ngokusebenzisa aws configure sso --profile . Bese, ungena usebenzisa iphrofayili edaliwe aws sso login --profile .

Cabangela ukucushwa okuzenzakalelayo kokuhlukahluka kwemvelo okuzenzakalelayo. Lokhu kufaka phakathi imvelo eguquguqukayo echazwe kulokhu okuthunyelwe, njengokuthi AWS_REGION, CLAUDE_CODE_USE_BEDROCK, ANTHROPIC_MODELfuthi ANTHROPIC_SMALL_FAST_MODEL. Lokhu kuzolungiselela Ikhodi ye-Claude ukuze ixhume ngokuzenzakalelayo i-Amazon Bedrock, inikeze isisekelo esingaguquki sokuthuthukiswa kuwo wonke amaqembu. Izinhlangano zingaqala ngokunikeza onjiniyela imiyalelo yokuzisiza.

Cabangela ukusebenzisa Isiqondiso se-Claude Code nge-Amazon Bedrock ukuthunyelwa kwamabhizinisi amakhulu. Umhlahlandlela usiza izinhlangano ukuthi zikhiphe Ikhodi ye-Claude kuyilapho zigcina ukulawula okuqinile kokufinyelela kwensiza ye-AI, ixhumeka kalula kungqalasizinda yobunikazi ekhona futhi ihlinzeka ngokubonakala kokukhiqiza konjiniyela namaphethini okusetshenziswa.

Buyekeza izabelo zesevisi futhi usethe amathokheni afanele ngomzuzu (TPM) kanye nezicelo ngomzuzu (RPM) ngokusekelwe enanini lonjiniyela abasebenzayo. Qiniseka ukuthi unezabelo ezanele ze-TPM ne-RPM ukuze usekele ukusetshenziswa kwethimba lakho. Kunconywa ukuthi usebenzise i-akhawunti eyodwa ye-AWS ukuze uthole inkomba. Ukuze uthole isiqondiso, landela izincomo zomkhawulo wesilinganiso. Isibonelo, uma unonjiniyela abangu-200, ungase ucele i-TPM engu-20,000 kunjiniyela ngamunye, noma izigidi ezingu-4 seziphelele ze-TPM.

Cabangela izimvume, inkumbulo, namaseva e-Model Context Protocol (MCP) enhlangano yakho. Amaqembu ezokuphepha angamisa izimvume eziphethwe zalokho i-Claude Code eyiyo nokungavunyelwe ukuyenza, engakwazi ukucishwa ukucushwa kwendawo. Ngaphezu kwalokho, ungakwazi ukumisa inkumbulo kuwo wonke amaphrojekthi ukuze ukwazi ukwengeza ngokuzenzakalela ukugeleza komsebenzi wamafayela omyalo we-bash, nezimiso zesitayela ukuze ziqondaniswe nokuthandwa yinhlangano yakho. Lokhu kungenziwa ngokuthumela eyakho CLAUDE.md ifayela kuhla lwemibhalo yebhizinisi //CLAUDE.md noma uhla lwemibhalo lwasekhaya lomsebenzisi ~/.claude/CLAUDE.md. Okokugcina, sincoma ukuthi ithimba elilodwa elimaphakathi lilungiselele amaseva e-MCP futhi lihlole a .mcp.json ukucushwa ku-codebase ukuze abasebenzisi bazuze.

Cabangela ukuhlola ukuphepha okuphelele okusekelwe ezidingweni zenhlangano yakho lapho uhlanganisa i-Claude Code, umkhiqizo wenkampani yangaphandle, ne-Amazon Bedrock. Ngenkathi i-Amazon Bedrock ihlinzeka ngezici zokuphepha eziqinile, i-Claude Code isebenza ngokuzimela emishinini yendawo, edinga ukuhlolwa okuhlukile nokulawulwa kobungozi. Cabangela ukusebenzisa izilawuli ezifanele nokuqapha Ikhodi ye-Claude, efana nendlela ongabhekana ngayo namanye ama-software ezinkampani zangaphandle asebenza endaweni yakho.

Isiphetho

Kulokhu okuthunyelwe, ufunde ukuthi ungaxhuma kanjani i-Amazon Bedrock ku-AWS GovCloud (US) Region nge-Claude Code. Le nhlanganisela inika amandla amandla okuhlela ama-ajenti ezimboni ezilawulwayo. Ngokusebenzisa ingqalasizinda yefu evikelekile ye-Amazon, i-Amazon Bedrock ukuze iqonde, kanye nomsizi we-AI othuthukisiwe we-Claude Code, izinhlangano manje zingakwazi ukucubungula ama-codebase abucayi ngokuphumelelayo ngenkathi zigcina ukuhambisana okuqinile nezidingo zobukhosi ngokusebenzisa i-AWS GovCloud. Ukwengeza, ngokusebenzisa ukugcinwa kwesikhashana okusheshayo, abasebenzisi bangabona izindleko ezincishisiwe kanye nezikhathi zokuphendula ezithuthukisiwe. Lokhu kuhumusha ngokuqondile ekusebenzelaneni okusheshayo, okungokwemvelo okwengeziwe ngekhodi kuzilungiselelo ezivikeleke kakhulu. Ngokuhlanganisa i-Amazon Bedrock Esifundeni se-AWS GovCloud (US) kanye nethuluzi lokufaka amakhodi le-ejenti njenge-Claude Code, amakhasimende manje ayakwazi ukufinyelela ezindleleni ezisabelayo, ezingabizi kakhulu, nezihlakaniphile ekuthuthukisweni kwesoftware evikeleke kakhudlwana ezimbonini ezilawulwayo. Abasebenzisi manje sebengakwazi ukusebenzisa amandla ohlelo lwe-AI ye-agent ngenkathi benamathela kumazinga aphezulu okuvikela idatha nokuhambisana.

Ukuze uthole ulwazi olwengeziwe:

I-Amazon Bedrock prompt caching ngokuvamile itholakala ngeClaude Sonnet 4.5. Ukuze ufunde kabanzi, bona i-Prompt caching ukuze uthole ukuchazwa kwemodeli esheshayo kanye ne-Amazon Bedrock.

Ikhodi ye-Anthropic Claude manje isiyatholakala. Ukuze ufunde kabanzi, bona uhlolojikelele lwekhodi ye-Claude futhi uxhumane nethimba lakho le-akhawunti ye-AWS ukuze uthole isiqondiso mayelana nokusetshenziswa.

Hlola i-Supercharge ukuthuthukiswa kwakho nge-Claude Code kanye ne-Amazon Bedrock esheshayo yokulondoloza isikhashana ngokusebenzisa amamodeli wakamuva Ezifundeni zezentengiso.

Mayelana nababhali

Bradley Wyman uyi-Solutions Architect eqenjini le-AWS Aerospace & Satellite, lapho esiza khona amakhasimende e-Aerospace & Satellite ukuthi asebenzise ubuchwepheshe obuphambili ukuze axazulule izinselele zebhizinisi eziyinkimbinkimbi. Ngentshiseko ejulile ye-Generative AI, i-Agentic Development, nokuthuthukiswa komsebenzi wesimanje, uBradley uletha ubuchwepheshe obukhethekile ekusizeni izinhlangano ukuguqula ukusebenza kwazo ngokusebenzisa izixazululo zamafu ezintsha. Uzinikele ekwenzeni ubuchwepheshe be-AI obuphambili bufinyeleleke futhi bube nomthelela kumakhasimende asemkhathini nawesathelayithi, okubenza bakwazi ukuphusha imingcele yalokho okungenzeka embonini yabo.

I-Doug Hairfield i-Senior Solutions Architect eqenjini le-AWS Aerospace & Satellite esiza izinhlangano ukuthi zisebenzise amandla e-AI ukuze zixazulule izinkinga zomhlaba wangempela. Uletha ulwazi olujulile lokusiza amakhasimende omkhakha womphakathi aklame imithwalo yawo yomsebenzi ezindaweni zokuthobela okuphezulu. Uma engaklami izixazululo zamafu, uzomthola engubaba wentombazane futhi ejabulela isikhathi nomndeni wakhe.

Keith Martin waqala umsebenzi wakhe wokuba unjiniyela we-software e-NASA Mission Control, lapho ukwethembeka kwesistimu ekusebenzeni kokundiza komkhathi kungaxoxiswana. Kusukela lapho usesebenzise leso siyalo ekwakhiweni kwezinhlelo zokuhweba emkhakheni wezamandla, eqondisa amakhasimende e-AWS Aerospace & Satellite nge-GovCloud kanye nokuthunyelwa kokuthotshelwa komthetho okuphezulu, futhi manje usehola izisombululo ze-automation ne-AI kuma-AWS Infrastructure Systems in Supply Chain Operations Excellence. Umbono wakhe nge-AI uqondile: iyithuluzi elilodwa phakathi kwabaningi, eliletha inani kuphela lapho lisekelwa izisekelo eziqinile zedatha kanye nezinjongo ezichazwe kahle.

Jonathan Evans ungumakhi Wezixazululo Zomhlaba Wonke we-Generative AI kwa-AWS, lapho esiza amakhasimende ukuthi asebenzise ubuchwepheshe be-AI obusezingeni eliphezulu ngamamodeli we-Anthropic's Claude ku-Amazon Bedrock, ukuze axazulule izinselele zebhizinisi eziyinkimbinkimbi. Ngesizinda sobunjiniyela be-AI/ML kanye nolwazi olusebenzayo olusekela ukuhamba komsebenzi komshini emafini, u-Jonathan unentshisekelo yokwenza i-AI ethuthukisiwe ifinyeleleke futhi ibe nomthelela ezinhlanganweni zabo bonke osayizi.