Indlela Yokuhlola Ukubonakala Komkhiqizo Kuzo Zonke Izinjini Ze-AI nge-EEO

Yini i-EEO?

EEO imele Everything Engine Optimization, okuyigama elihlanganisa i-SEO, i-AEO, i-GEO, ne-LLMO. Le phrojekthi isiza amabhrendi aqonde futhi athuthukise indlela avela ngayo ezimpendulweni ezikhiqizwa izinjini ze-AI.

I-EEO ihlanganisa amakhono amathathu ahlobene:

• Ukuhlolwa kwe-EEO: ibuza izinjini ze-AI imibuzo abangase bayibuze abathengi bangempela, bese ikala ukubonakala kwebhrendi ezimpendulweni ezibuyisiwe. Imibiko ihlanganisa nezimpendulo zangempela ukuze zihlolwe.
• Ukubhaliswa kwebhrendi: kugcina amakhadi ebhrendi afundeka yimishini alandela i-schema ethi eeo.brand.v1. Amabhrendi angafaka izimangalo ngamakhadi awo ngokuqinisekisa isizinda ku-pull request yomphakathi.
• Umakhi wesayithi: uguqula amakhadi ebhrendi abe yisayithi elimile eline-llms.txt, i-JSON-LD, kanye nefayela le-robots.txt elivumela abacaca be-AI kuphela.

Le phrojekthi yakhelwe ukuba ivuleke, inganciki kumadependency lapho kungenzeka khona, futhi ihlolwe kalula. Imithetho yayo ebanzi kanye nesivumelwano sokuziphatha esinemithetho emihlanu sabahlinzeki bokwenza i-optimization kubhalwe ku-standard ye-EEO.

Kungani ukubonakala kwebhrendi ye-AI kubalulekile

Uma umthengi ecela injini ye-AI isincomo, impendulo yakhiwa ngolwazi imodeli elufundile, njengezindatshana, izibuyekezo, namakhasi okufakwa ohlwini. Incazelo yebhrendi uqobo ingase ingafinyeleli kumodeli ngendlela ehlelekile.

I-EEO ibhekana nale nkinga ngokunikeza amabhrendi ifomethi evulekile yokubeka amaqiniso awo, inqubo yomphakathi yokuqinisekisa ubunikazi besizinda, namathuluzi enza ulwazi olutholakele lufundeke yizinjini, onjiniyela, nabathengi.

Khetha indlela yokusebenza ye-EEO

Indawo yokugcina ikhodi inikeza izindlela eziningana zokuqala. Khetha leyo ehambisana nenhloso yakho:

• Sebenzisa i-eeo-local.html ukuze uhlole ngokushesha kusiphequluli ngaphandle kokufaka lutho.
• Sebenzisa i-cli.cjs ukuze wenze ukuhlola okuphindaphindekayo, ukuhlolwa kwamaqoqo, ama-dataset, kanye nokuhlolwa kabusha okuhleliwe.
• Sebenzisa i-Docker image ukuze usebenzise isayithi lomphakathi lamabhrendi endaweni yangakini.
• Sebenzisa i-server.cjs kanye nephrojekthi ye-web/ ukuze usebenzise uhlelo lwewebhu oluzisingathayo lwethuluzi lokuhlola.
• Sebenzisa i-GitHub Action, iseva ye-MCP, noma isandiso se-Chrome lapho uhlanganisa ukuhlolwa nezindlela zokusebenza zokuthuthukisa noma zocwaningo.

Yenza ukuhlolwa kwefayela elilodwa

1. Dawuniloda futhi uvule ithuluzi

Dawuniloda i-eeo-local.html endaweni yokugcina ikhodi bese uyivula kusiphequluli. Lolu hlelo aludingi i-Node.js, ukufakwa kwamaphakheji, noma iseva.

2. Khetha izinjini futhi unikeze okhiye

Khetha izinjini kukhathalogi, equkethe ama-preset angu-26, okuhlanganisa izinjini ezingu-11 zomdabu wase-China. Namathisela okhiye be-API obudingwa izinjini ezikhethiwe. Ithuluzi ligcina laba khiye ku-localStorage yesiphequluli.

3. Gcwalisa ifomu lebhrendi

Faka ulwazi lwebhrendi bese ukhetha imboni efanele. Bese wenza ukuhlolwa. I-EEO ibuza imibuzo egxile kumthengi futhi ihlola ukuthi ibhrendi ivela ngokugqama kangakanani ezimpendulweni.

4. Hlola futhi ukhiphe umbiko

Umbiko uhlanganisa ibanga le-S ukuya ku-D, amaphuzu okuqwashisa nawokukhulunywa ngawo, ukuhlukaniswa ngokwenjini, kanye nazo zonke izimpendulo zangempela. Ungakhipha umphumela njenge-JSON noma uwuphrinte ube yi-PDF ukuze wabelane ngawo futhi uwubuyekeze.

Yenza ukuhlola usebenzisa i-CLI

Indlela yokusebenza yomugqa womyalo idinga i-Node 18 noma entsha. I-CLI ayinaso isinyathelo sokufaka ama-dependency endleleni yayo eyinhloko.

1. Dala ifayela lokumisa

Kopisha ukumisa kwesibonelo bese ugcwalisa izilungiselelo zezinjini nokhiye be-API ohlela ukubasebenzisa:

cp config.example.json config.json

Gcina ukumisa kuhambisana nendawo yakho yasendaweni futhi ungafaki izimfihlo endaweni yokugcina ikhodi yomphakathi.

2. Hlola ibhrendi eyodwa

Yenza ukuhlolwa kwebhrendi eyodwa ngokunikeza igama lebhrendi nemboni:

node cli.cjs check 蜜雪冰城 --industry 餐饮

I-CLI ithumela izicelo zokuhlola ezimisiwe futhi ikhiqize imiphumela ehambisanayo yokubonakala.

3. Yenza ukuhlolwa kweqoqo

Faka amagama amabhrendi efayeleni lombhalo, elilodwa emugqeni ngamunye, bese ubhala imiphumela ibe yi-CSV:

node cli.cjs batch brands.txt --csv result.csv

Lokhu kuyasiza uma uqhathanisa amabhrendi amaningi noma udala inqubo yokubika ephindaphindekayo.

4. Hlola i-dataset

Sebenzisa umyalo oqhutshwa yi-dataset lapho ufuna ukucubungula amabhrendi avela esigabeni semboni:

node cli.cjs dataset 白酒 --limit 10

Inketho ethi --limit igcina ukuhlolwa kwesibonelo kukuncane. Sebenzisa node cli.cjs --help ukuze ubone yonke imiyalo etholakalayo, okuhlanganisa ne-watch yokuhlolwa kabusha okuhleliwe.

Yenza isayithi lomphakathi lamabhrendi nge-Docker

Indlela ye-Docker yakha isayithi lebhrendi kusuka ku-datasets/brands-1k.json. Ayidingi okhiye be-API ngoba isebenzisa ulwazi lwebhrendi olukhiqiziwe, esikhundleni sokwenza ukuhlolwa kwenjini.

docker build -t eeo-community .
docker run -p 8080:80 eeo-community

Ngemva kokuqala kwe-container, vula i-http://localhost:8080. Isayithi linikeza ikhasi elilodwa ngebhrendi ngayinye futhi lihlanganisa ukusekelwa kwe-llms.txt, i-JSON-LD, kanye ne-robots.txt kwabacaca be-AI.

Lokhu kumiswa kwe-Docker kuhlukile kuseva yokuhlola ezisingathayo. Kwiseva yokuhlola, hlola i-server.cjs kanye ne-config.example.json. Isixhumi esibonakalayo sewebhu singakhiwa nge:

cd web
npm install
npm run build

Isixhumi esibonakalayo sewebhu esakhiwe kusengaphambili sifakiwe futhi ku-public/.

Qonda ukwakheka kwendawo yokugcina ikhodi

• eeo-local.html iyithuluzi lokuhlola elisefayeleni elilodwa elinama-preset enjini angu-26.
• cli.cjs inikeza izinqubo ze-check, batch, dataset, ne-watch.
• core.cjs iqukethe injini yokuhlola eyabiwe esetshenziswa yi-CLI, iseva ye-MCP, ne-GitHub Action.
• server.cjs iwuhlelo lweseva oluzisingathayo, kanti web/ iqukethe umthombo wayo we-React ne-Vite.
• action/ iqukethe i-GitHub Action yokusebenzisa ukuhlolwa kokubonakala komkhiqizo ku-CI.
• mcp/ iqukethe iseva ye-MCP ye-Claude Desktop, Cursor, ne-Windsurf.
• extension/ iqukethe isandiso sokuhlola ngokushesha se-Chrome MV3.
• datasets/ iqukethe iqoqo eliphelele lemikhiqizo nedatha yokuqala ekhethiwe.
• platform/ iqukethe imibhalo yokuqinisekisa izimangalo neyokuqedela.
• standards/taxonomy/ iqukethe izici zezimboni namabhange emibuzo yabathengi.
• tools/ iqukethe abakhiqizi bamasayithi, llms.txt, ama-schema, nama-index.

Faka isicelo noma wengeze ikhadi lomkhiqizo

Faka isicelo ekhadini elikhona

Imikhiqizo kufanele iqale ngokusesha datasets/ ukuze ithole ikhadi elikhona. Inqubo yokufaka isicelo ichazwe ku-platform/README.md, futhi isebenzisa ukuqinisekiswa kwesizinda nge-pull request yomphakathi.

Le nqubo inikeza irekhodi elibonakalayo lokuthi umkhiqizo ulawula isizinda esiqinisekisiwe esihlotshaniswa nekhadi.

Faka umkhiqizo ongekho

Uma umkhiqizo ungafakwanga ohlwini, dala bese uthumela ikhadi ngokuvumelana ne-CONTRIBUTING.md. Landela i-schema nemithetho yokunikela yale repository ngaphambi kokuvula i-pull request.

Izinqubo ezithuthukile namathiphu okunikela

Xhuma ama-endpoint e-AI ahambisanayo

Ababukeli bezinjini bangaphakamisa ama-preset amasha enjini ekhathalogi yokuthola. I-CLI, iseva ezisingathayo, neseva ye-MCP nakho kungaxhuma kuma-endpoint ahambisana ne-OpenAI nge-config.json. Buyekeza iziqondiso zokunikela ngaphambi kokwengeza i-preset noma ukushintsha indlela yokuthola.

Yenza ukuhlolwa okuphindaphindwayo ngokuzenzakalelayo

Sebenzisa umyalo we-watch we-CLI ukuze wenze ukuhlolwa kabusha okuhleliwe. Ekuzenzakalelweni kwe-repository, ifolda ye-action/ inikeza i-GitHub Action engasebenzisa ukuhlolwa kokubonakala ku-CI. I-repository iphinde ibe nezinqubo zokuqinisekisa izimangalo, ukuqedela izimangalo, nokuzihlola kwanyanga zonke.

Sebenzisa iseva ye-MCP kumathuluzi e-AI

Iseva ye-MCP iyatholakala ku-Claude Desktop, Cursor, ne-Windsurf. Iphinda isebenzise injini yokuhlola eyabiwe, okwenza amaklayenti ahambisanayo akwazi ukufinyelela emandleni okuhlola ephrojekthi ngokuhlanganiswa kwe-MCP. Yilungiselele usebenzisa izilungiselelo ezijwayelekile zenjini nezikhiye ezichazwe ku-config.example.json.

Nweba ama-taxonomy ezimboni

Abanikeli be-taxonomy bangangeza imitapo yezici zezimboni namabhange emibuzo yabathengi ngaphansi kwe-standards/taxonomy/. Imithetho yokwamukela ibhalwe ku-taxonomy README. Ukwengeza i-alias yomkhakha kuchazwe njengomnikelo wokuqala olula.

Khiqiza izimpahla zokushicilela ezifundeka yimishini

Amathuluzi okukhiqiza amasayithi angaguqula amakhadi emikhiqizo abe isayithi lolwazi elingashintshi. Umphumela okhiqiziwe uhlose ukwenza amaqiniso emikhiqizo kube lula ukuthi izinjini namanye amathuluzi awathole nge-llms.txt, i-JSON-LD, nemithetho yabacrawler.

Gcina ukuhlolwa kuqondakala

Imibiko ye-EEO igcina yonke impendulo yoqobo yenjini esikhundleni sokuveza amaphuzu kuphela. Lapho uqhathanisa izikhathi zokusebenza, buyekeza ukuhlukaniswa kwenjini ngayinye nezimpendulo zomthombo kanye namaphuzu okuqwashisa nawokubalulwa. Lokhu kwenza kube lula ukuhlukanisa ushintsho olubanzi ekubonakaleni kumphumela odalwe yinjini eyodwa noma umbuzo owodwa.

Ukuze uthole imiphumela ephindaphindekayo, gcina ukukhethwa komkhakha, idatha yomkhiqizo, ukucushwa kwenjini, nokufakwa kwe-batch kufana phakathi kwezikhathi zokusebenza. Lapho ungeza injini noma i-taxonomy entsha, bhala phansi lolo shintsho ukuze imibiko yakamuva iqhathaniswe ngobulungisa.

Isiphetho

I-EEO inikeza inqubo evulekile yokukala indlela izinjini ze-AI ezichaza ngayo imikhiqizo kanye nokushicilela amaqiniso emikhiqizo ngendlela ehlelekile. Qala nge-eeo-local.html ukuze uhlole ngaphandle kokufaka lutho, dlulela ku-CLI ngomsebenzi we-batch noma ohleliwe, bese usebenzisa i-Docker noma iseva ezisingathayo lapho udinga isibonelo somphakathi esisebenza endaweni.

Uma usuqonda inqubo eyisisekelo, ungafaka izicelo zamakhadi emikhiqizo, unikele ngedatha ye-taxonomy, wengeze ama-preset ezinjini, uhlanganise ukuhlolwa ku-CI, noma uxhume iseva ye-MCP kumathuluzi okuthuthukisa e-AI ahambisanayo. Iphrojekthi ikhishwa ngaphansi kwe-Apache License 2.0, kanti amafayela e-standard nawe-taxonomy asebenzisa amalayisensi abhalwe ku-repository.