The iterative development of the Internet and digital technology is like a war drum urging the public to demonstrate the possibility, richness, challenge, and speculation of Internet literature and art again and again. It also brings us a core question: Internet literature and art as a spiritual way for people to grasp the world.
In March, the warm and humid air currents from the ocean began to become stronger. At the same time, the cold air that has been entrenched for a winter is at the end of its strength. After several "contests" between cold and warm air, it has brought frequent precipitation to many places in our country.
It is cloudy and rainy every day in northern Guangxi, and the sunshine in southern Guangxi is "online for a limited time and in a limited amount". The Guangxi Meteorological Observatory predicts that in the next week, there will be a lot of shortwave trough and weak cold air activity, and low-level shear lines and stationary fronts will swing in northern Guangxi. Light rain will occur every day in northern Guangxi, with occasional moderate rain "blind boxes" in local areas. Southern Guangxi will be cloudy to overcast with scattered light rain.
In terms of temperature, in the next week, the average temperature will be 1°C to 2°C higher in northwest Guangxi and 2°C to 4°C higher in other areas. In terms of rainfall, rainfall in Baise, Hechi and the northern parts of Guilin and Liuzhou cities was 30% to 80% higher, and locally the rainfall was more than double, while other areas were 40% to 90% lower.
The temperature in Guangxi gradually rises, showing the characteristics of "warm during the day and cool in the morning and evening". From the 18th to the 21st, the maximum temperature in most parts of Guangxi was 23°C to 27°C; from the 22nd to the 23rd, the temperature in parts of western and southern Guangxi was 27°C to 28°C, and locally it would soar to 29°C to 32°C. Warm weather is coming.
Before the 20th, most of the lowest temperatures were still hovering around "1", and the temperature difference between morning and evening was obvious. Everyone should pay attention to adding or removing clothing in a timely manner.
The Nanning Meteorological Observatory predicts that from the evening of the 17th to the day of the 18th, urban areas of Nanning will be overcast with localized light rain, with temperatures ranging from 19°C to 24°C and southerly winds of magnitude 2 to 3. Occasionally there will be light rain, so be sure to bring rain gear when traveling.
According to the national spring mask map launched by China Weather Network, the temperature fluctuations in Guangxi, Hainan and other places are generally not large. The highest temperature in some places may reach above 20°C. The body feels warm and comfortable, and there is no need for spring mask. You can flexibly adjust your outfit according to your own conditions.
Nanning Media Center Reporter Zhao Jinling/Wen Song Yankang/Photo Guangxi Meteorology Nanning Meteorology
When it comes to English, how many children frown and look embarrassed? How many parents are looking sad and helpless?
Memorizing and forgetting words, struggling to read texts, not daring to speak in class, frequently holding back on test scores, the more children learn, the more resistant they become, and even become irritable when they see an English book, gradually falling into a vicious cycle of "fear of learning - tired of learning - falling behind". As parents, we feel sorry for our children's frustrations and are also anxious about their prospects for further studies. We always want to find a shortcut to help our children get rid of their English difficulties and regain their confidence in learning.
In fact, it’s not that children don’t love English by nature, it’s just that they haven’t encountered the right method or the right classroom. This summer vacation, don’t force your children to memorize by rote, and don’t let English become the trigger of parent-child conflicts. Qisu English Summer Camp is coming, using 17 years of brand accumulation and 31 camps of practical experience to help children completely break their psychological barriers and achieve a gorgeous transformation from "fear of English" to "love English", turning English from a "stumbling block" into a "plus point"!
Say goodbye to boring memorization and use interesting methods to ignite your interest in English
Traditional English learning relies entirely on mechanical repetition, memorizing words in isolation, and mechanically memorizing grammar. The boring process will only wear away children's patience and enthusiasm. Qisu English Summer Camp directly addresses the pain points of learning, relying on the national intellectual property certified mind map + story word recording patent method to completely subvert the traditional learning model, turn boring words into interesting stories, and connect scattered knowledge points into a clear knowledge network.
We tailor learning plans for primary and secondary school students at different learning stages: junior high school students use 200 mind maps, covering 1,800 core vocabulary words for the high school entrance exam, with thematic classification and logical connections. Children build a vocabulary system like playing puzzles, and memorizing words is no longer a burden; high school students use 24 original bilingual stories to understand 3,500 high-frequency words for the college entrance examination. Science fiction, adventure, growth and other diverse themes allow children to follow the stories to learn English, unknowingly accumulate vocabulary, improve their sense of language, and become more addicted as they learn.
Three teachers accompany students and escort them throughout the process, eliminating the fear of learning.
Children's fear of English often stems from classroom pressure and frustration. Qisu English Summer Camp implements a three-teacher learning model to give children a full sense of security and belonging. The main teachers are all senior teachers with junior high school certificates and have been engaged in English in primary and secondary schools for many years. They are well versed in the psychology of teenagers and are good at using interactive and heuristic teaching to create a relaxed and pleasant classroom atmosphere so that children dare to speak and express themselves. The assistant teachers provide personal guidance throughout the process, patiently correcting pronunciation, answering questions, checking for gaps one-on-one, and helping children overcome their weak links. The life teacher takes care of the food and daily life, and protects the safety of the children 24 hours a day, so that the children can devote themselves to learning without worries.
The camp adopts fully enclosed and meticulous management, with 24-hour security patrols, professional nutritionists custom-made meals, and full-time medical staff on duty to fully protect children's learning and life. Parents can trust them with peace of mind, and children can grow up happily.
Closed-loop anti-forgetting system, turning interests into reality and improving scores
Many parents are worried: If they have a lot of fun during the camp period, will they return to their original shape after the camp ends? Qisu English’s original 7311 anti-forgetting closed-loop system completely solves the problem of forgetting. 7 days of immersive training to break through vocabulary difficulties, 30 days of online review and consolidation, 100 days of on-the-job tracking coaching, and 1 year of full-time tracking companionship. Together with the AI intelligent learning system, exclusive study plans and online refresher training, the children's learning effects can be continuously consolidated and truly achieve short-term breakthroughs and long-term benefits.
There is also a 3-year tuition-free retraining guarantee, so children can return to the camp at any time to strengthen their shortcomings, so that their English improvement will not be abandoned halfway, and their interest will be transformed into lasting motivation for learning.
Real transformation is visible, and countless children have regained their confidence in English
Since the 31st session of the camp, Qisu English has witnessed the growth of countless children: children who were originally resistant to English and had less than 1,000 words, mastered core vocabulary in 7 days, and took the initiative to raise their hands to read the text after the start of school; children who were introverted and dared not speak, became confident and generous in the oral interaction at the camp, and their English scores steadily improved; students who were serious in subjects relied on efficient learning methods to make up for their shortcomings, and successfully counterattacked and became the top English students in the class.
The change in children's attitude towards English is the most gratifying gain for parents. Here, children gain not only improved scores, but also the courage to challenge, the ability to learn independently, and English thinking that will benefit them for life.
This summer vacation, don’t let your children continue to worry about English, and don’t let their resistance delay their growth prospects. Qisu English Summer Camp uses fun teaching to ignite interest, uses professional teachers to protect you, and uses hard-core effects to witness transformation. It truly allows children to let go of their fears, fall in love with English, and easily start their English counterattack!
Early bird registration is currently open, and limited-time discount seats are limited. Hurry and lock in growth opportunities for your children. This summer, wait for your children to emerge from their cocoons and become butterflies, and learn English with a new look!
The iterative development of the Internet and digital technology is like a war drum urging the public to demonstrate the possibility, richness, challenge, and speculation of Internet literature and art again and again. It also brings us a core question: Internet literature and art are a spiritual way for people to grasp the world. How should we recognize and grasp it? The author tries to make a superficial discussion from the perspective of artistic language.
Any art has its own special artistic language. Art is not only the physical presentation of aesthetics, but also the mediated presentation of artistic language. Literature, music, art, and dance are all the same. So, does online literature and art have its own unique artistic language? If not now, will it happen in the future?
It seems a bit premature to raise this issue now. Its premise is to assume that Internet literature and art has become an independent art category, but in fact, at present, both Internet literature and Internet audio-visuals are only "a rich totality" composed of various innovative types relying on the Internet as a medium, and their independence has not yet been fully, significantly, and clearly demonstrated.
However, according to McLuhan's theory, from print media to electronic media, and then to digital media, changes in the media environment and ecology have a profound impact on the renewal of aesthetic experience. The aesthetic methods and aesthetic activity characteristics formed by Internet thinking have increasingly appeared in various innovative types of online literature and art. Online movies are watched in slices, online dramas are watched at double speed, short videos are disseminated socially, and the reading experience of online literature is also very different from traditional reading. These increasingly normalized, universal, and popular phenomena make us feel that whether there is inherent uniqueness in the artistic language of online literature and art seems worthy of our serious consideration and objective answer.
The first thing that needs to be seen is that the development and evolution of artistic forms are deeply and closely connected with the underlying logic of the media. Media is an extension of human senses. In traditional art, the mediated presentation of its artistic language also manifests itself as the concentrated strengthening of certain human sensory functions in aesthetic experience. For example, music and radio drama are to the auditory sense, calligraphy and art are to the visual sense. Among traditional arts, stage is a relatively "comprehensive" art category, which includes literature, rap, instrumental music, performance and many other artistic elements to varying degrees. However, its comprehensiveness is limited to the time and space of the aesthetic scene and does not have the extended function of media. Only after entering the era of electronic media did "comprehensive art" break through the limitations of time and space. New comprehensive arts such as film and television also created a new artistic language after realizing the extension of media.
Digital media is a new iteration compared to electronic media. The Internet is not only the communication medium for all types of online literature and art, but Bit is also the physical medium for all types of online literature and art. In the world of bits, the extension of the Internet media to people's senses is no longer singular, but comprehensive, three-dimensional, and integrated. Artificial intelligence is regarded as an extension of the human nervous system. This kind of media extension integrates the real world and the virtual world and creates comprehensive interaction. It is a two-dimensional media immersion, rather than the one-dimensional media immersion of the virtual world created by movies and TV. Therefore, the degree of "comprehensiveness" and expression of online literature and art in art, compared with the comprehensive art produced in the era of electronic media such as traditional film and television, have completely new characteristics, occupy a new level of magnitude, and are showing increasingly clear aesthetic differences.
"Comprehensive art" will gradually shape its own unique artistic language as it matures, and will eventually be familiar and accepted by the public. For example, the artistic language of traditional opera is often unified by "program" in the industry. The artistic language of movies and TV is often referred to as "shot" or "montage". Can online literature and art become a new type of "comprehensive art"? In recent years, many experts and scholars have begun to discuss its aesthetic characteristics and artistic language issues. Based on the underlying logic of Internet media, they have successively proposed a series of discourse concepts such as "interactive immersion", "scene" and even "body" as the possibility of naming the aesthetic language of Internet literature and art. Of course, this discussion still needs to wait for the next development and practice of online literature and art to gradually become clear.
What also needs to be noted is that different media have different epistemological biases. Even if the stories of novels and movies are the same, they still convey to people two completely different artistic "reality" and aesthetic experiences. In his famous theory of media evolution, Levinson regarded the Internet media as a new technological stage that "reshapes reality." The Internet has re-corrected the sensory bias strengthened by previous media technologies and the "partial reality" lost thereby. While breaking the limitations of human physiology and realizing the extension of media, it has returned to the "complete reality" of full sensory experience. The accelerated application of virtual reality technology and artificial intelligence has brought us an extremely intuitive experience.
Today, in the process of imitating and copying people's perception and cognitive models, technology has moved from "mediatization" to "de-mediation", that is, the progress of media technology has reached a level where people cannot feel the existence of media. This will inevitably have a subtle impact on the development of artistic forms. For example, spoken language is the earliest medium born by human beings. Folk storytelling, rap and other arts that rely on spoken language have a strong on-the-spot nature in their initial stages. After the advancement of media caused oral expression to be replaced by written expression, sensory bias led to changes in aesthetic patterns. The reason why a large number of colloquial expressions have reappeared in online literature is not only because of the non-professional nature of the authors. The deeper reason may also be that the "de-mediation" of the Internet has brought people back to the "live" aesthetic experience mode.
Of course, the most typical manifestation of "de-mediation" at the moment is the superimposed application field of virtual reality technology and artificial intelligence. Corresponding to online literature and art, this field will bring new, full-sensory experience, more perceptual and intuitive aesthetic objects, as well as the expansion of the boundaries of artistic creativity under the model of human-machine co-creation and integration. The coexistence of virtual reality and real reality, and the continued evolution of the cooperation between artificial intelligence and humans, will further enhance the "comprehensive" and "de-mediation" characteristics of online literature and art, thus profoundly affecting the development trend of its artistic language.
The Marxist view of literature and art believes that literature and art are people’s grasp of the “artistic spirit” of the world, an intuitive and perceptual grasp. "De-mediation" provides more convenient conditions for this kind of intuitive and perceptual life experience, but in the final analysis, the "humanization" of media is still a higher stage of technology in the field of online literature and art to create people's aesthetic consciousness and accumulate aesthetic experience. Only people can master the world in a spiritual and aesthetic way, and only people. Therefore, the development of the artistic language of online literature and art is still based on the laws of human aesthetic spiritual activities and further deepens the aesthetic relationship between people and the world. It’s worth waiting and watching what kind of new “language” it will produce in the future.
(The author is a doctoral candidate in Media Arts at Communication University of China and deputy director of the Internet Literature and Art Communication Center of China Federation of Literary and Art Circles)
After living in Japan for five years, many people will come to an extremely pessimistic conclusion: this country has no future. However, this view often only sees the silence of the surface, but ignores the unfathomable resilience of the underlying layer. When we want to discuss a country's future, we cannot rely on intuition; we must see through its cold and hard-core data.
First of all, please put away your misconceptions about Japan’s “poverty”. Even after experiencing the so-called "lost thirty years", Japan's per capita nominal GDP still remains at US$35,000. If you include the huge income from overseas assets (GNP), this number is actually US$38,000. More importantly, Japan is an extremely "average" country, with the median annual income of ordinary people ranging from 150,000 to 200,000 yuan. Among all G7 countries, Japan's tax system does not encourage sudden wealth. It is more like a "socialist country" disguised as capitalism. Although this system has stifled the enthusiasm of a very small number of geniuses, it has built an extremely thick anti-risk cushion for the entire society.
Let’s talk about the automobile industry, which everyone loves to pessimize. Yes, the rise of China's electric vehicles is indeed unstoppable, but has Japan really lost? The profits of Toyota alone exceed the combined profits of all new energy vehicle manufacturers in China. In terms of stability and quality control of internal combustion engines, Japan still holds absolute hegemony. New energy is a big gamble, and Japan has chosen to bet long on hybrid, hydrogen and traditional energy. This kind of "conservatism" seemed outdated in the era of fanaticism, but when the energy crisis broke out, it became the most stable safe haven.
What is even more terrifying is Japan’s “invisible hegemony”. In the fields of semiconductor upstream materials, high-precision equipment and robots, Japan has already deeply embedded itself in the marrow of the global industrial chain. It does not seek to compete with the United States in terms of originality from 0 to 1, but in terms of ultimate craftsmanship from 1 to 100, it is the "gatekeeper" of the global industry. Looking at Japan's five major trading companies, they account for 20% of Japan's output value. In the past thirty years, they have used extremely cheap Japanese yen financing to harvest minerals, farmland and core assets around the world, recreating an "invisible Japan" overseas.
Of course, Japan's terminal illness lies not in external competition, but in its internal "rigid paradox." Its aging has penetrated deep into its bones, and more than 30,000 century-old companies are facing the dilemma of having no successors. Last year, hundreds of companies were forced to close down simply because they could not recruit workers. An extremely absurd logic emerges here: Japan's economy is desperate for foreign labor and global talent, but Japanese society is permeated by an extremely narrow-minded xenophobic sentiment. This tear between demand and cognition is Japan’s real fatal wound.
But we cannot conclude that Japan has no future. Japan is experiencing a "deep precipitation" similar to Thailand. Thailand seemed to be stagnant after the 1998 financial crisis, but it has developed world-class standards in the fields of cultural creativity, design and advertising. The same is true for Japan. It still has strong defense capabilities in terms of hardware support and cultural industry output in the AI era.
Therefore, Japan is not "dying", it is just transforming from an "expansionary empire" to a "defensive fortress." It has a profound heritage, a very high social lower limit, and an irreplaceable industrial status. Its future may no longer see turbulent growth, but it has more “dry food” reserves in the dark than most countries. Looking at the bottom of data and logic, Japan's current situation and structure tell us: you can laugh at its slowness, but you absolutely cannot underestimate its ability.
Thanks to the IT House netizen who is very homely and afraid of strangers for submitting clues!
IT House reported on March 17 that ASUS Fearless 16 notebook 2026 notebook has now opened new product reservations on JD.com, with optional Ryzen 7 8845H / Intel Core Ultra 9 285H processors, priced from 4,199 yuan.
IT Home attaches the details of each new product as follows:
Fearless 16 Ryzen Edition 2026 144Hz:
This notebook is equipped with AMD Ryzen 7 8845H second-generation AI processor, with 8-core 16-thread design, built-in Radeon 780M integrated display, body thickness of 15.9mm, weight of 1.65kg, performance release up to 45W, AI computing power up to 38TOPS, maximum acceleration frequency up to 5.1GHz, built-in 16GB LPDDR5X memory, 512GB PCIe 4.0 The solid-state drive uses dual M.2 slots and supports expansion up to 4TB.
In addition, the notebook is equipped with a 70Wh large-capacity battery, supports USB-C charging, has a full-size backlit keyboard, is equipped with a 16:10 144Hz high refresh screen, covers 100% sRGB color gamut, supports Windows Hello infrared face recognition login, and is priced at 4,199 yuan.
JD.com Asus Dreadnought 16 Ryzen Edition 4199 Yuan direct link to Dreadnought 16 Ryzen Edition 2026 2.5K 144Hz:
This notebook is also equipped with AMD Ryzen 7 8845H processor, 16GB of memory, 1TB PCIe 4.0 solid-state drive, dual M.2 slot design, 45W performance release, support for adjusting three-speed heat dissipation mode, all-metal body design, equipped with 2.5K 144Hz screen, covering 100% sRGB color gamut, maximum brightness of 400nit, and supports USB-C charging.
In addition, the machine is equipped with dual USB-A 3.2, dual USB-C 3.2 and HDMI 2.1 interfaces and a 3.5mm audio interface. It is equipped with a full-size backlit keyboard, a 7.1-inch Big Mac touch version, and is priced at 4,999 yuan.
JD.com ASUS Dreadnought 16 Ryzen Edition (standard pressure R7 16G 1T) 4999 yuan direct link to Dreadnought 16 second generation Core Ultra9:
This notebook is equipped with Intel Core Ultra 9 285H standard voltage processor, with a 16-core 16-thread design, a turbo frequency of up to 5.4GHz, an AI computing power of up to 99TOPS, a built-in 8Xe high-power core display, a built-in 16GB onboard memory that can be added up to 48GB, and a 1TB PCIe 4.0 solid-state drive.
In addition, the machine is equipped with a 2.5K 144Hz gaming-grade screen, covering 100% sRGB color gamut, with a maximum brightness of 400 nits. It has a new generation ergonomic keyboard, the direction keys are half-height design, and is equipped with a 6.1-inch large-size touch panel with a thickness of 1.69cm. It has two full-function USB-C, two USB-A and an HDMI interface. It is priced at 5,999 yuan.
JD ASUS Dreadnought 16 Second Generation Core Ultra9 5999 Yuan direct link
In 2025, the field of AI crawlers will usher in new changes. This article focuses on the best practices of AI crawlers in 2025, and provides an in-depth practical demonstration of how to use the three major tool combinations of Deepseek, Crawl4ai, and Playwright MCP to achieve efficient and intelligent crawler operations. From environment construction to code practice, to dynamic loading and data extraction, it fully demonstrates the charm and potential of AI crawlers, allowing you to easily master cutting-edge crawler technology.
Let’s do a practical exercise today: use Crawl4ai to make an AI crawler and see what it looks like.
Test your skills
First install it according to the official code:
#Install the package
pip install -U crawl4ai
# Run post-installation setup
crawl4ai-setup
#Verify your installation
crawl4ai-doctor
Finally, when you see the picture below, it proves that the installation and initialization were successful.
Next, let’s test the official example.
import asyncio
from crawl4ai import *
async def main:
async with AsyncWebCrawler as crawler:
result = await crawler.arun(
url=”https://www.nbcnews.com/business”,
print(result.markdown)
if __name__ == “__main__”:
asyncio.run(main)
The official test is a news website
Create a new py file, paste the code into it, and run it directly. The results show that it can indeed be crawled normally.
Extract form
The latest version of Crawl4ai has a new function: after crawling the tables on the website, parse them into pandas DataFrame format.
To put it simply, before we needed to manually clean and structure the downloaded data, then convert it into DataFrame format for analysis.
Now you can do it in one step.
Let's take a look at this official example for a virtual currency website. We need to crawl down the table in the picture below and convert it into a python table, which can be directly used for the next step of analysis.
But there is a problem here: the official example cannot be used, as shown in the picture, it is incomplete, all are red wavy lines, and an error will be reported when running directly. My coding ability is poor and I can’t change it. What should I do?
It's very simple, just let AI change it. Go directly to Cursor.
But then there is a new problem: because this is a new function of crawl4ai, some AI should not have learned.
At this time, you can use the context7MCP we introduced before to let the AI learn the latest documents by itself and then complete the code. Red and warm! Cursor writes random code again? Install Context7 MCP in 1 minute to enjoy real-time document retrieval service
Prompt words I use:
The file code is the official example of crawl4ai. The effect is to capture the website table data as shown in the picture and save it in pandas dataframe format. However, this code is incomplete. You need to use context7 mcp to find the latest crawl4ai document and complete the code to ensure that it can be used normally.
After the AI operates for a while, the code it gets can be run directly. We see that the form from the previous website has been successfully downloaded into DF.
It went quite smoothly. I opened Excel to see it more completely, and then I could use the data for analysis.
dynamic loading
Today's websites are rarely static. Most of them are dynamically loaded, which means they need to be scrolled continuously to load new content. It would be too troublesome to handle this process by yourself.
Fortunately, Crawl4ai has built-in javascript support. We can directly write a js code to load all the content in one scroll of the page. result = await crawler.arun( url="https://dynamic content site.com", js_code="window.scrollTo(0,
document.body.scrollHeight);", wait_for="document.querySelector('.loaded')")
OK, so far, we have run through the official crawler sample code provided by Crawl4ai. But AI has not been used yet.
You know, the reason why we use these frameworks is to let AI help us solve the problems in crawling.
So let’s take a look at how to use AI to make crawlers in Crawl4ai? Advanced: Dynamically load large models to crawl e-commerce reviews
The business scenario comes first. I chose a high-frequency scenario: crawling e-commerce product reviews. (The comment data that is subsequently crawled can also be subjected to text analysis to unearth information of commercial value)
The URL is:
https://www.amazon.com/PawSwing-AutoComb-Automatic-Surround-biomimetic/dp/B0DMSVNTC1
Scroll down to see the list of comments:
Utilizing playwright MCP initialization script
In the original Cursor window, let AI help us write the code first: Now you need to write a Crawl4ai script to grab the comments under this Amazon product: customer name, title, country, time, comment content, etc. You can use playwright mcp to check the website first, and then modify it.
I didn’t find it. I didn’t let AI write the code directly. Instead, I asked it to take a look at what the website looked like and then write the code.
Because the loading process, speed, and structure of each website are different, if you rush to write a general code, it may not work.
As for the introduction and installation of Playwright MCP, I have said it before. You can jump directly to this article to learn: Use Playwright MCP to let AI change the shit mountain code it wrote.
Having said that, we can already see the AI Called MCP tool automatically opening the Amazon website and `get_visible_html`, that is, taking a look.
The obtained code is as follows. It is very long overall. I have cut out some key parts, including suggestions, and put them in comments:
1. Define the data model for Amazon reviews
classAmazonReview(BaseModel):
customer_name:
review_title:
country_and_date:
review_body:
image_urls:
rating:
# 2. Use LLMConfig to configure the AI model
llm_config = LLMConfig(
provider=provider,
api_token=api_token,
base_url=base_url
# 3. Set the strategy for AI crawling data. The key is the prompt word.
strategy = LLMExtractionStrategy(
llm_config=llm_config,
schema=AmazonReview.model_json_schema,
extraction_type=”schema”,
instruction=f”””
Extract Amazon product review information from the provided HTML content.
Reviews are usually contained in a hook with the 'data-hook="review"' attribute
element.
Please extract the following information for each comment and construct it into a list of JSON objects:
1. `customer_name`: The name of the reviewer, usually within or near a span element with 'data-hook="genome-widget"'.
2. `review_title`: The title of the review, usually within a span or a element with 'data-hook="review-title"', which may be bold text.
3. `country_and_date`: The country and date of the review, usually within a span element with 'data-hook="review-date"', in a format similar to "Reviewed in on".
4. `review_body`: The body content of the review, usually within a span element with 'data-hook="review-body"'.
5. `image_urls`: List of image URLs uploaded by users in comments. Pictures are usually
tag, its parent element may have a 'review-image-tile' or similar class. Please extract
The 'src' attribute of the tag. If there are no images, this field is null or an empty list.
6. `rating`: The star rating of the review, usually near the review title or at the beginning of the review body.
Make sure to extract as many reviews as possible. Ignore content in non-comment areas on the page.
""",
chunk_token_threshold=8000, # Change back to a reasonable chunking threshold
apply_chunking=True, #Chunk large pages
input_format="html",
verbose=True# Enable detailed logs of LLM policy
# 4. Browser settings must be set, especially for websites with strong anti-crawling capabilities.
browser_config = BrowserConfig(
headless=False, # can be set to True to run in the background
java_script_enabled=True, # Ensure JavaScript loads comments
# You can add proxy, user-agent and other configurations to simulate real users and reduce the risk of being blocked.
# user_agent=”Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36″
viewport={“width”: 1280, “height”: 800},
verbose=True# Enable detailed logs of browser configuration
# 5. Crawler configuration
crawler_config = CrawlerRunConfig(
cache_mode=CacheMode.BYPASS,
page_timeout=90000, # Increase the page loading timeout, Amazon pages may be slow
extraction_strategy=strategy,
# Increase the waiting time to ensure that dynamically loaded comment content appears
# Note: crawl4ai currently does not have a direct wait_for_selector or a fine-grained waiting mechanism similar to playwright
# You can control the waiting time indirectly through page_timeout, or consider using playwright to operate directly later.
verbose=True# Enable detailed logs of crawler operations
# 6. Start crawling
result = await crawler.arun(
url=self.url,
config=self.crawler_config,
js_code="window.scrollTo(0, document.body.scrollHeight);",
wait_for="document.querySelector('.loaded')"
Complete crawling
Look at the result, it’s pretty good
In particular, image_url can directly organize image addresses.
The logic of Crawl4ai is that it will first throw all the HTML to AI and then let AI come up with a parsing strategy, and then split it into multiple modules and give them to AI to parse the data one by one.
Here’s a key point: When choosing a model, pay attention to one that supports “big context”
The deepseek v3 I chose at the beginning was too small, so I couldn’t read the entire HTML at first and reported an error.
Later, I switched to gemini-2.5-pro-exp-03-25 and succeeded.
It’s a waste of tokens. You can feel it. I grabbed 24 comments and how much was consumed:
Total consumption: 992,830 (Prompt) + 6,348 (Completion) = 999,178 tokens
if
Calculating the cost of gemini-2.5-pro-preview-03-25, it is about 1.3 US dollars (Gemini helped me calculate it. Fortunately, I am using the free version.)
Summarize
To be honest, RPA is much more effective when it comes to crawlers, but RPA seems to be a false proposition. You need to set up processes, capture elements, design capture logic, etc. A codeless tool will make novices confused and have to learn from scratch.
The more feasible solution for AI crawlers now is based on Cursor and paired with Playwright MCP to develop Crawl4ai scripts.
Although it is not very easy to use now, it may be a very smooth AI crawler experience in the near future.
The Automotive Aftermarket Is Huge, And Individual Stores Urgently Need To Improve Their Supply Chain And Management Capabilities.2026-03-22Comments(0)
Jiang Xin Cheated Liu Tao, Is The Entertainment Industry So Cruel Or Brainless? How Did This Friendship Change?2026-02-27Comments(0)