{"id":58007,"date":"2026-10-14T09:57:08","date_gmt":"2026-10-14T01:57:08","guid":{"rendered":"https:\/\/www.1ai.net\/?p=58007"},"modified":"2026-09-20T15:03:13","modified_gmt":"2026-09-20T07:03:13","slug":"ai%e6%8f%90%e7%a4%ba%e8%af%8d%e5%88%9b%e4%bd%9c%e7%ac%ac%e5%85%ad%e5%8d%81%e4%b8%80%e8%8a%82%ef%bc%9a%e4%b8%8d%e7%94%a8%e6%ad%bb%e7%a3%95%e9%95%bf%e6%8f%90%e7%a4%ba%e8%af%8d%ef%bc%9a","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/58007.html","title":{"rendered":"Al-Ai warns section 60: super-simple codex! You don't have to use a long hint. You do it with Codex"},"content":{"rendered":"

It's the first time many people have seen it Codex<\/a>\"Software to help programmers write code.\" But for ordinary creators, it is more worthwhile to learn how to deal with tasks: read information, understand objectives, call tools, implement steps, check results, and then revise them on the basis of feedback\u3002<\/p>\n

After this section, you should be able to organize a vague idea into a missionable list and do a different set of short plays, video clippings, small games, etc. in the same way\u3002
\n
\nI. UNDERSTANDING CODEX ' S METHODS OF WORK<\/strong><\/p>\n

Why does the same sentence, \"Do me a short play,\" that ordinary chat tools might give you only a piece of text, while Codex can continue to read information, create files, call tools and check results<\/p>\n

Because Codex's focus is not just on answering questions, but on carrying out tasks. A full mission usually consists of five layers:<\/p>\n

Target: What exactly is this time to do, for example, to produce an ancient wind drama\u3002<\/p>\n

Source: Where are the role maps, scripts, videos, tables or project documents placed\u3002<\/p>\n

Fixed methods: Rules by which character design, script structure, spectrometry and hints should be made\u3002<\/p>\n

Implementation tool: Documents, commands, browsers, editing software or connected office tools are needed\u3002<\/p>\n

Acceptance and inspection criteria: which documents were eventually delivered and which errors were checked\u3002<\/p>\n

In this case, regular mission message explains \u201cwhat to do this time\u201d; and Skill saves \u201cwhat to do with such tasks in the future\u201d. In the absence of Skill, you have to re-interpret the character card, the three-story structure, the camera chart and the verb format every time you do a short play; and, with Skill, these fixed rules will only have to be organized once, followed by the addition of the subject, role and duration\u3002<\/p>\n

here's what we'll need for this time<\/p>\n

skill is not a magic button, but a reusable instructions. it can consist of steps, rules, reference documents and scripts that are suitable for recurring work such as short plays, production of commodity maps, organization of interviews, examination of lessons, etc\u3002<\/p>\n

II. Installation of Codex<\/strong><\/p>\n

Click this link to open the official network:<\/p>\n

https:\/\/openai.com\/codex\/<\/p>\n

\"93263<\/p>\n

Click to Download<\/p>\n

III. STUDY IN G.C.S.D<\/strong><\/p>\n

\"c8ebadaaj00tli3pf006vd000v90hkp\"<\/p>\n

A stable Codex mission can be expressed in four parts:<\/p>\n

G, GOAL GOAL GOAL: THE END OF WHAT\u3002<\/p>\n

C,Context context: which documents, pictures, videos or references need to be read\u3002<\/p>\n

S, Steps \/ Skill Steps and Skills: What skills should be mobilized, what rules should be followed and what process should be followed\u3002<\/p>\n

D,Done completes the standard: to output what files, check what questions and what status is over\u3002<\/p>\n

The four need not be written in tables or long. Their role is to reduce Codex's guess space\u3002<\/p>\n

\"a94d077ej00tli3pp005ed000v90lp\"<\/p>\n

A \u201chigher\u201d \u201cbetter rhythm\u201d cannot be directly accepted. Replace them with the \" 60-second vertical screen \" , the \" no repeat lens \" \" , the \" output project file and film \" , the \" 60-second vertical screen \" , the \" no repeat lens \" , for the mission to be truly operational\u3002
\n
\nIV. Nine types of practical play for Codex<\/strong><\/p>\n

\"58f0a462j00tli3q100ud000v90lhp\"<\/p>\n

The nine types of game appear to be different, with the bottom being access to information, tools, missions and inspection results\u3002<\/p>\n

These games appear to be very far-reaching, but at the bottom they are the same: Codex reads the information first, then calls the tools allowed in the current environment, performs the task and checks the results. Distinguishing is only about input data, tools and acceptance standards\u3002<\/p>\n

1. AUTO-GENERATED AI SHORTS<\/p>\n

It is appropriate to move a story idea forward to character-setting, scripts, spectroscopy and mirror-by-simulation tips. You need to provide subject matter, time, role reference and style; Codex calls the writer or director, Skill, to produce intermediate results in fixed steps. Don't just ask for a final piece, first check on the character card and the lens. The post will be completely dismantled\u3002<\/p>\n

2. Automatic clipping of video<\/p>\n

Put all raw materials, bystanders and reference pieces in the folder, and then write the theme, duration, scale and rhythm. Mr. Codex can make a list of material, select candidate clips and then call on the available editing tool to finish the rough cut. The black frame, sound break, repeat camera and subtitle mask are checked during the acceptance, rather than just \"Does it export video?\"\u3002<\/p>\n

Auto Generate Game<\/p>\n

It starts with a description of the core game, how it works, art style and customs or business cycle. Codex does a minimum playable version, actually runs and then adds tasks, stores, archives and role systems based on feedback. Warm Valley Journal is a full example of this game, and the post will show the real picture\u3002<\/p>\n

4. Website development and App<\/p>\n

Starting with a line of business objectives, you can gradually generate page structure, interface styles, interactive logic and data functionality. For example:<\/p>\n

i looked for a picture for ui-style reference<\/p>\n

and enter:<\/p>\n

\"Let's think of an APP that can record how to play cards<\/p>\n

codex used image 2.0 to generate this UI interface<\/p>\n

Then direct codex to create an actionable APP interface<\/p>\n

5. Mass management local material<\/p>\n

It is suitable for the organization of hundreds of pictures, videos or documents: reading document information, sorting by item, uniform naming, finding duplicate items, generating a list of material. The security practice is to allow Codex to export the \"Plan Change List\" and the \"Suspect Duplication List\" before moving or deleting them\u3002<\/p>\n

6. CLEANING OF THE C DISC<\/p>\n

The word \u201cclean-up\u201d here should be understood as analysing occupancy and assisting with safe handling, rather than allowing Codex to remove system files at will. It can measure large files, caches, download directories and duplicate installation packages, first giving a list of capacity and risks\u3002<\/p>\n

i don't know<\/p>\n

ANALYSES C DISK SPACE OCCUPANCY, LISTS THE LARGEST FOLDERS AND CANDIDATES FOR SAFE CLEAN-UP. IN THE FIRST PHASE, ONLY REPORTS ARE GENERATED AND NOTHING IS DELETED. THE SYSTEM CATALOGUE, SOFTWARE RUNNING DOCUMENTS, PERSONAL ITEMS AND UNCERTAIN DOCUMENTS ARE KEPT\u3002<\/p>\n

7. Operation of the Flying Book Office<\/p>\n

When a flight book tool or browser has been connected and has been given the appropriate permission, Codex can organize the document, the matrix box, generate a meeting record, update the task or write the information on a given page. The task must clarify the workspace, the target document and the range of permitted changes\u3002<\/p>\n

8. Automatic installation of software and tools<\/p>\n

Codex can first check what's in the computer and then select the right version to complete the download, installation, configuration of environmental variables and running validation\u3002<\/p>\n

9. Production exclusive Skill<\/p>\n

When the same set of tasks is repeated two or three times, the method can be sealed into Skill. For example, \u201creading oral scripts, refining teaching objectives, producing illustrations, writing curricula, examining web pages\u201d has been made a regular process. In the future, Codex will be able to implement the same standard by providing new material\u3002<\/p>\n

For example, I'm the one who made the curriculum<\/p>\n

I can get Codex to summarize it into a skill<\/p>\n

Case I: Upload three Skills, one sentence to generate an AI short play<\/strong><\/p>\n

This is actually so complicated. You don't have to write character cards, scene sheets and scripts in turn, and you don't have to direct each step alone. The real operation has only three things: uploading three Skill files, telling story ideas in one sentence, asking Codex to press Skill to do it automatically\u3002<\/p>\n

\"fa5ba14aj00tli3rz004rd000v9000g0p\"<\/p>\n

You're in charge of creativity. Three Skylls are in charge of scripts, characters and scenes<\/p>\n

Three, Skill<\/p>\n

Skill: Responsible for the subject, audience, character relations, the skeleton of the story, a catalogue and a full script. Simplely understood, it answers: \u201cWhat does this short play say?\u201d\u3002<\/p>\n

Person Skill: Responsible for identity, age, quality, clothing, makeup, face, film sense and graphic video style. It simply understands that it answers \u201cwhat a person should look like\u201d\u3002<\/p>\n

3x3 scene Skill: in charge of space such as residences, atoms, study rooms, courtyards, export of buildings, displays, proprietors, light and nine-magic scenery tips. Simple understanding, it answers: \"Where is the story?\"\u3002<\/p>\n

Copy links to web site paste, download automatically<\/p>\n

https:\/\/superi.mfdemo.cn\/chaoji\/2026\/06\/15\/6a2fd45506ea2.rar<\/p>\n

How does White actually work<\/p>\n

opens a new task, uploads three .md files together, then gives a general idea:<\/p>\n

\"b44ee5d2j00tli3sg000zd000lv00atp\"<\/p>\n

prompt:<\/p>\n

based on these three skills, work with each other to give me a 90-s short play from this story<\/p>\n

She was forced to marry her sister into the General's house. The young general is said to have returned in cold blood and cold blood, and to have returned from the border, to have been killed and not close to her\u3002
\nBut on her wedding night, she found the general with a poison and was dying. In order to save her life, she pretended to be a good wife and treated him secretly\u3002
\nThe general initially suspected that she was a conspirator. It was later discovered that this seemingly weak little doctor could save lives and kill people\u3002
\nCore: Bridalships, Medical Women, Coldface Generals, Mutual Tests, Days and Days\u3002<\/p>\n

AI WILL GIVE YOU A SET OF SCRIPTS. YOU CAN TALK ABOUT CHANGES\u3002<\/p>\n

\"ad8cd5dfj00tli3s000td000lx009ep\"<\/p>\n

WHEN THE SCRIPT IS SATISFACTORY, THE HINT IS USED TO ALLOW AI TO GENERATE CHARACTERS AND SCENES (EMPHASIS IS PLACED ON A WHITE MAKE-UP PHOTOGRAPH)\u3002<\/p>\n

\"e2694297j00tli3t50092d000v90nvp\"<\/p>\n

when the hint is generated, you can direct him to the image 2.0<\/p>\n

\"062796c1j00tli3te008kd000m201b0p\"<\/p>\n

THEN I ASKED AI TO GIVE A DETAILED STORY AND A SCRIPT\u3002<\/p>\n

\"75d91b48j00tli3tm000wd000lq00qxp\"<\/p>\n

With a spectroscopy script, AI will generate a set of video alerts that are appropriate for soeance 2.0\u3002<\/p>\n

i don't know<\/p>\n

based on the spectroscopy, give me a video message that matches the seedance 2.0<\/p>\n

\"c53ce701j00tli3u7009bd000v90ihp\"<\/p>\n

AND LET AI GENERATE VIDEO-TEMPORARIES AND PALETTES:<\/p>\n

prompt:<\/p>\n

>and give me a picture of the atmosphere that matches the women's frequency and a colored hint. >\u3002<\/p>\n

\"72db31f5j00tli3ug007kd000v900i4p\"<\/p>\n

That's how the video alerts are generated<\/p>\n

then open low https:\/\/labs.google\/fx\/tools\/flow<\/p>\n

use nanobanana pro to generate a third view of the person. why don't you use the image 2.0 to create it? 'cause image 2.0 isn't good enough, it's not real enough, it's oily<\/p>\n

Enter the hint:<\/p>\n

The three-face view of the person, the three-view view of the person in the picture, the white-floor map, and the one on the left, which shows the half-person portrait of the character. Its facial features, skeletal structure, dementia, hair and head dressing must be strictly consistent with the nature of the skin texture, hair holes and hairline details. On the right side, the full-body picture of the role is presented in a standard A-pose position, organized in three precise views: positive view, side view and back map. Supermodel body size, complete clothing design, fabric level, shoes and overall dress style must be strictly followed. All clothing and accessories present a real sense of physical material quality\u3002<\/p>\n

\"e9fe49cdj00tli3uw005vd000v90aiip\"<\/p>\n

\"275<\/p>\n

And then you turn on the dream to generate a video<\/p>\n

All you have to do is upload a third view of the person, a scene, and then enter a hint (air colour + video)\u3002<\/p>\n

\"7697f098j00tli3vd00e0d000tc00ahp\"<\/p>\n

This video contains the following:<\/p>\n

The old woman's sense of emotional shortness, the real reality of the movie, the east of the world, the beauty, the danger of confusion, restraint and love after marriage. The focus of the picture has always been on changes in the eyes of men and women, in breathing distance, in hand contact, in clothing sleeve friction and in micro-expressions. The male lords have a sense of restraint, protection and oppression, but are not gruesome evils; the female lords are sober and wise, with a soft middle, and remain active in close proximity\u3002
\n9:16 for the stand-up, with priority given to a two-person peri-view, a side peri-view, eye and hand feature. The Zhu red veil, Hippa, screen edges are used as a cover for the future, creating an implicit vision and spatial layer. The person's face is clear, the background is modestly debilitated, the light is deep, the light is soft and purified, and the real lens breathes. It's very sophisticated, but it's not stylish. It's like the key scenes of an ancient play\u3002
\nReal and natural Eastern faces, with skin texture and facial bones, soft, but not peeling, and wet eyes with no tears. Women ' s makeup is clean, low saturated, classical, and men ' s makeup is clean and cold. Real silk, knitting, sarcasm, deep wood and cold iron. Clothes have weight, and silk and rout are used only to limit their movement\u3002
\nThe light uses both warm-coloured candlelight and cold and blue moonlight. The warm and bright face of the person and the red dress, and the cold light glittering of the male master ' s black shirt and sense of danger. The face is in a soft and dark transition, with clear eyes and a clear vision to avoid mass death. The red veil, candle-fired scenes, small amounts of cigarettes and air dust add to the film level. The whole body is low in saturation, low in light, warm in warm contrast, expensive and not glamorous, sweet and unsatisfied\u3002
\nThe motion of the lens is smooth and follows with slow advance, light side movement, small rounding and stability. The action scene was short and clear, and the emotional scene slowed down, leaving a standstill and a visualization. The image is of a high-quality, old-fashioned scenery, without the use of a fairy filter, a net-red skin, a low-cost studio red filter or excessive moisture\u3002<\/p>\n

do not produce any subtitles, text, text. do not generate any bgm, but keep sound
\nPhoto 1 Su's night make-up photo: locks on Su's late face, hair, Zhu's wedding dress, green underwear, silver needle hair and deep-blown medicine bag\u3002
\nPhoto 2 Shao-seong ' s make-up photograph: locking down Xiao-seong ' s ashes with face type, beaming hair, black General ' s uniform, dark gold nails, cold iron wrists, black and blue lines and swords\u3002
\nPicture 3 of the scenes of the marriage: locking down the wedding house of the General, the Zhu's red-hyeon screen, pulling the bed, the side window, candlelight and cold moonlight\u3002
\nSniper 01, pic. 3@Marriage scene. The very nearness is fixed, the Zhu's hippa is naturally down and full of images, and the fabric is clear. A long sword with a small amount of dry blood slowly enters from below the scene, slamming at the edge of the Hippa and holding back. The background is fading and only warm candles are kept. The sound is the sound of a slight arcade impact and a low sting from a distance\u3002
\nMirror 02 is based on photo 1 Sou's night makeup, photo 2 Shaw's sundown makeup and photo 3's wedding scene. At night, Sue covered the Zhu's Hippa, sitting in front of the happy bed, with a close view on his face, and his right hand in his wide sleeve, with a silver needle in his fingertips. Shaw's ashes entered the horizon only in black clothes and boots and stopped in front of her. Sue stayed still at night and his fingers were slightly tightened. Cameras are stable, shallow. The sound effects are sank feet, clothing frictions and quiet breathing. Mirror 03
\nUse three pieces of material to keep people in line with clothing. The low-altitude feature follows the long sword, the stunned Hippa slowly with the sword, and the red veil falls naturally, displaying a calm and vigilant face. Swords go up, gently hold her jaw. The camera moves slowly, highlighting metal, red veils and eyes. The sound of a blade rubbing a bead. Mirror 0-4
\nReference to three pieces. The two sides were near, the plume was leaning towards the night, and the sword was holding her underside, and their faces were close. Sue didn't come back late and looked up to him. The camera is moving very slowly, capturing the sense of oppression and limiting ambiguity. Xiaoxie said, \"Sou Jin has a mole in his right hand. You don't.\" The background is just a slightly shaken red veil and candlelight. Mirror 05
\nReference to three pieces. Su's subjective view of the night, with the prospect of the sword debilitated, the focus shifted steadily to the wrist of the sword held in the ashes of Shaw, and a black and blue venom line that spreads under the cold-iron wrist. The focus then goes back to the sobering eyes of Sue. Su said in the evening and quiet: \u201cThe General will examine the bride on the night of the wedding or the postmortem?\u201d Precise the details and do not use a sharp mirror\u3002<\/p>\n

Case two: Put the material together, get Cordex to manipulate the mouse<\/strong><\/p>\n

It's also direct: put the video material in the same folder and tell Codex, \"Do me a rough cut,\" it looks at the material and operates the editing software with computer control\u3002<\/p>\n

\"86a1fefbj00tli3w5003ed000v90eup\"<\/p>\n

What do you mean, \"crunch\"<\/p>\n

A rough cut is not the end of the picture. It mainly introduces materials into the software, removes the apparently ineffective parts, and sets the remaining segments around the time line. Subtitle special effects, music card points, fine mosaics and emotional rhythms usually require manual processing\u3002<\/p>\n

How does White actually work<\/p>\n

Put the material in one folder<\/p>\n

\"79b2e4c1j00tli3wd008nd000oy00g2p\"<\/p>\n

just talk to codex<\/p>\n

Read all the video material in the current folder and help me finish a rough clip. Please open the editing software and import the material and schedule it to the time line by mouse and keyboard. Delete the black screen, the apparent loss of focus, the repetition and the total futility. Do not delete or overwrite raw material. Stop at the clip time line after completion, let me check\u3002<\/p>\n

\"24865fc5j00tli3wl0022d000v90nlp\"<\/p>\n

Codex, how do you actually cut it<\/p>\n

Codex opens the editing software, clicks on the import, browses the folder, drags the material, moves the time line and executes the cut like a person. It's not that the editing software suddenly has a universal button, but Codex is replacing the mouse and keyboard\u3002<\/p>\n

It's important to make clear the real shortcomings<\/p>\n

At present, the size of the mouse is very slow and the effect is more general\u3002<\/p>\n

Step by step to identify interfaces, clicks and drags, much later than professional editors\u3002<\/p>\n

The more material, the longer it takes to view and select\u3002<\/p>\n

Visible black screens, repetitions and invalid clips can be handled, but with limited judgement of emotions, laughter points, card points and advanced rhythms\u3002<\/p>\n

The mouse drag may not be accurate when the time line is complex\u3002<\/p>\n

When Windows operates the desktop software, the front mouse keyboard is usually used\u3002<\/p>\n

So it's good for a first version of a rough cut or a show that Codex can manipulate a computer, not for a single key to get a professional film\u3002<\/p>\n

Case III: Paste a hint, Codex directly produces the Warm Valley Journal<\/strong><\/p>\n

Operational step: Paste a complete hint to have Codex interview you first; once the need has been identified, it will create, run, test and fix files directly in the current working area\u3002<\/p>\n

\"45bb845cj00tli3xb003cd000v90eup\"<\/p>\n

First a few questions, then Cordex takes the initiative<\/p>\n

The full message is actually used:<\/p>\n

I'd like to make a little game like Starbucks and Cereal
\nYOU'RE A SENIOR INDEPENDENT GAME DEVELOPER, GAME DESIGNER AND UI DESIGNER. PLEASE WORK WITH ME TO MAKE A LITTLE GAME FROM ZERO THAT CAN ACTUALLY RUN AND PLAY\u3002
\nI have only one preliminary idea at the moment, and you need to help me refine it by asking questions and then finish it directly in the current working area\u3002
\nWorkflow:
\nPhase one: Needs interviews, please ask me one or three of the most important questions, and continue to follow up on my answers. At a minimum, the following should be recognized:
\n3. Operating platforms such as browsers, Windows, MacOS or mobile phones
\nDon't start with a lot of questions. Where I'm not sure, please provide two or three specific options and recommend a formula that is most suitable for quick completion\u3002
\nAfter the interview is completed, please organize:
\n\u2013 Introduction to a Word Game \u2013 Core Play Cycle \u2013 Operating Method \u2013 Victory, Failed or Growing Conditions \u2013 Functional List \u2013 Technical Scheme \u2013 Fine Arts and Sound Schemes \u2013 What's not done in the first edition
\nIF MY VISION IS TOO BIG, FOR EXAMPLE, TO RECAPITULATE STARBUCKS, PLEASE TAKE THE INITIATIVE TO REDUCE TO A COMPLETE, FUNABLE MVP, BUT RETAIN THE MOST INTERESTING CORE EXPERIENCE\u3002
\nPhase III: Start development
\n3. Create and modify the required documentation. 5. Prioritize the use of simple and reliable realization, avoiding unnecessary complexity
\nPhase 4: Tests and grinding
\n\u2013 Start the game and actually test \u2013 Check the console error \u2013 Test start, pause, fail, win and restart process \u2013 Check the display under common screen size \u2013 Fix the problems found \u2013 round grinding the image, feedback and operating hand feeling
\nFinally tell me:
\n\u2013 What are the functions of the game \u2013 how to start and play \u2013 where are the main files \u2013 what are the limitations of the current version
\nCooperation rules:
\n\u2013 Ask questions when needs are not clear, do not start production without permission \u2013 do not wait for my approval at every step after needs are confirmed \u2013 do not stop at the planning stage and complete the operational version \u2013 For every important stage completed, give me a short progress statement -- if you have a technical problem, investigate and try to solve it yourself -- control the size of the first release, first \"complete, fun, run\" and then consider adding something
\nPlease start with the interview. Ask me first and foremost\u3002<\/p>\n

Then you'll be able to answer the questions on a continuous basis, the question phase, and you can also answer the images of the style you want to refer to\u3002<\/p>\n

\"2718bfbj00tli3xw00ild000mk01ftp\"<\/p>\n

WAIT FOR AI TO GENERATE ITSELF\u3002<\/p>\n

WAIT 10 MINUTES BEFORE THIS IS THE GAME INTERFACE CREATED BY AI\u3002<\/p>\n

\"95f7e23cj00tli3y4002zd000v90ctp\"<\/p>\n

it's a bit of a stylish time, and it's time to upload your reference, and then call image 2.0 to generate the picture\u3002<\/p>\n

prompt:<\/p>\n

now the ui and the picture are too simple, and i've given you several reference maps, using the image 2.0 model, to generate the assets needed for the game. including ui, figures, maps, tools, houses, seeds, seedlings, etc<\/p>\n

\"895a252ej00tli3yn00a6d000m300kzp\"<\/p>\n

And that's the most important way that Codex makes a little game: instead of asking you to write code, it's going to talk about it with an interview, and then try to finish it and tell it where it needs to be changed\u3002
\n
\nConclusion<\/strong><\/p>\n

Academia Codex is not the same thing as going back to more terms, nor is it the same thing as writing a long hint. A truly useful ability is to organize ideas into implementable, verifiable and retrofittable tasks\u3002<\/p>","protected":false},"excerpt":{"rendered":"

For the first time, many people saw Codex and understood it as \"software to help programmers write code.\" For ordinary creators, however, it is more worthwhile to learn how to deal with tasks: read information, understand objectives, call tools, implement steps, check results and then revise them on the basis of feedback. After this section, you should be able to organize a vague idea into a missionable list and do a different set of short plays, video clippings, small games, etc. in the same way. I understand how Codex works, why does the same sentence, \"Do me a short play,\" that ordinary chat tools might give you only a piece of text, and Codex can continue to read, create, call and check the results? Because Codex's focus is not just on answering questions, but on performing<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[149,144],"tags":[802,3149,956,5321,6623,8649,192,6481],"collection":[],"class_list":{"0":"post-58007","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"hentry","6":"category-jiaocheng","7":"category-baike","8":"tag-ai","12":"tag-codex","14":"tag-prompt","15":"tag-6481"},"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/58007","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=58007"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/58007\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=58007"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=58007"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=58007"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=58007"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}