Introduction

Hello everyone, I’m Leng Yi. Today I will be testing Kimi’s latest full-stack assistant, “OK Computer.”

A few days ago, I received a thank-you letter from Kimi, expressing gratitude for my support last year through tips. They offered me membership benefits equivalent to my tip amount, allowing me to experience the latest model capabilities firsthand.

Honestly, I didn’t expect Kimi to remember its early supporters. This was really thoughtful! So, I didn’t hesitate to upgrade to their highest tier, Moderato membership (which allows 20 uses of “OK Computer”) and received an additional 5 months of membership.

After upgrading, I quickly received an invitation to the internal testing of “OK Computer” (all users who tipped would receive priority invitations).

Experience with OK Computer

At the Yunqi Conference a few days ago, I saw an interesting product called “AI Exchange” (a platform that connects AI demanders and developers). I decided to see if I could use “OK Computer” to create a website prototype.

1) Developing the AI Exchange Website

I opened Kimi’s official website, kimi.com, selected “OK Computer,” and it was ready to use (I could also see my usage quota).

I input the task:

Project Template: Build an online trading platform for AI demanders (buyers) and AI developers (sellers), providing a secure and efficient matching mechanism, supporting the release, purchase, negotiation, and display of AI-related services, products, and agents.

Functional Requirements

1. User System Registration and Login: Support phone number registration. User roles: demander (Buyer) and developer (Seller), which can be concurrent. Personal center:
   • Buyer: Demand management, transaction records, favorite services.
   • Seller: Service/model release, pricing management, transaction records.
2. AI Service/Product Release and Display: Developers can publish services by filling in service descriptions, function scope, prices (fixed price/bargaining), and delivery cycles. Display page: Service detail page: functions, prices, case studies, rating leaderboard/recommendation page: display based on popularity, ratings, and transaction volume. Search and filter: by price, tags, AI fields (e.g., voice, image, text, video), delivery cycles.
3. Demand Release and Matching: Buyers can publish clear demands (e.g., “I need an image recognition agent with a budget of 2000 yuan”). The system recommends suitable sellers, or sellers can bid actively.
4. Transaction System Matching Logic: Supports bargaining and fixed-price direct orders. Payment process: funds escrow, released after delivery confirmation. Order management: status transitions (pending confirmation → in development → pending delivery → completed/canceled).
5. Credit and Evaluation: Completed orders allow buyers to rate and evaluate sellers. The platform displays developers’ credit ratings and transaction history.
6. Display and Recommendations: Homepage sections: popular demands, quality developer recommendations, recent transaction displays. Dynamic wall: real-time scrolling of the latest transactions. Case library: showcases quality successful cases.

First, let’s look at the finished product.

Experience URL:
https://vcnj4jhe2thpy.ok.kimi.link/index.html

The overall functionality is quite complete, and it’s the website prototype I wanted. How did it do this?

Once I sent the task, Kimi quickly powered up (the virtual computer) and got to work.

The first thing it did was act as a project manager, analyzing the entire requirement and breaking the project down into 11 sub-tasks.

Next, it continued as a product manager and UI designer, writing the PRD (Product Requirement Document) and visual design plan, clarifying the website features and visual design.

Since our website had a high demand for images, Kimi searched for relevant image materials and even generated a background image by itself. It created a resource folder and downloaded everything.

Then, it transformed into a front-end development engineer, developing HTML pages, including the homepage (index.html), service market page (marketplace.html), demand release page (demands.html), and personal center page (profile.html).

Finally, before deployment, Kimi acted as a testing engineer and operations engineer, conducting final functionality tests and optimizations before deploying to the server.

During its first check, “OK Computer” found a terminal runtime failure. It tried a new port and ultimately succeeded in deployment.

The final link was delivered to us, accessible publicly and shareable with others, viewable on both mobile and desktop.

Using the same prompt, I ran it again, this time with a more modern tech feel.

Experience URL:
https://kuleem2nugt64.ok.kimi.link/

2) Pixel Art Interview Program

Next, I had Kimi run a more complex task.

The prompt was:

Project Objective: Create a complete pixel art web application simulating a Western TV news/music interview program themed “Coldplay Concert Kiss-Cam Incident and Public Privacy Discussion”, including 3 minutes of dual audio and 20 synchronized pixel art images. —#

Visual Style Requirements

• Overall Style: 8-bit pixel art + elements from Western TV news/concert broadcasts (live broadcast graphics, news tickers)
• Color Scheme: Retro game colors (#FF6B6B, #4ECDC4, #45B7D1, #96CEB4)
• Character Design: 2 pixel characters
  • Host (Western news anchor style)
  • Guest (media/cultural commentator style)
• Background Elements:
  • Coldplay concert stadium stage
  • Large audience area (glow sticks, mobile screens)
  • Pixelated Kiss-Cam large screen framing
  • Pixelated social media interface
  • Studio commentary scene

Audio Content Requirements

• Duration: 3 minutes (180 seconds)
• Language: English (news podcast style)
• Format: Dual conversation
• Theme: Discussing the privacy issues raised by the Coldplay concert kiss-cam incident, including social media dissemination, brand reactions, and future outlook.
• Structure:
  • 0–30s Opening: Host introduces the background of the incident
  • 30–90s: Incident dissemination chain (live screen → audience recording → social media)
  • 90–150s: Subsequent reactions (company investigation, artist response, fan culture)
  • 150–180s: Future outlook (privacy reminders, concert management, platform responsibilities)
• Character Setting:
  • Host (Anchor): Calm, professional
  • Guest (Commentator): Media/sociological analysis, explaining how the incident became a global topic

Image Generation Requirements

• Quantity: 20 pixel art illustrations
• Size: 320×240 (retro game console resolution)
• Switching Frequency: Every 9 seconds, synchronized with audio
• Content Types:
  • Character Portraits (6 images): Different expressions and poses of the host/guest
  • Scene Illustrations (8 images): 1. Studio scene (news anchor desk) 2. Stadium night scene (Coldplay stage lights, glowing audience) 3. Kiss-Cam pixelated screen (crowd pixelated mosaic) 4. Close-up of fan area (waving glow sticks) 5. Social media interface (pixelated tweets/comments) 6. Company meeting room (silhouette style) 7. Coldplay stage background (lights and confetti) 8. News broadcast graphic (“Privacy Debate”)
  • Data Visualization (6 images):
    • Popularity curve, retweet volume bar chart, dissemination chain illustration, privacy risk matrix, fact-checking process, future improvement checklist

Prompt Template:
[pixel art], [8-bit retro game style], [Western TV news broadcast + stadium concert scene], [bright retro colors], [320×240 resolution], [no real faces recognizable]—#

Web Functionality Requirements

1. Custom pixel art audio player
2. Audio and image timeline synchronization (switch every 9 seconds)
3. Pixel UI control panel (play, pause, speed, subtitle toggle)
4. Responsive design (desktop & mobile, maintaining pixel clarity)

Technical Implementation Plan

Step 1: Audio Generation

• Generate a 3-minute dual conversation in English using AI voice synthesis (two voice tones: Anchor/Commentator)
• MP3 format, 128kbps, 44.1kHz
• Script divided into 4 segments (every 30 seconds for synchronization)

Step 2: Image Generation

• Use pixel art models to generate 20 images, ensuring consistent color and style across stadium, stage, and news studio elements

Step 3: Web Development

• Tech stack: HTML5 + CSS3 + JS
• Use Audio API to synchronize image carousel
• CSS3 pixel animations (fade-in, flashing subtitles)

Compliance and Narrative Boundaries

• Do not display recognizable private faces, only use mosaics/back views to represent the audience
• Focus discussion on public events, privacy issues, and cultural responses
• Add a footer statement on the webpage:
  “This is a pixel-art simulation for educational and creative purposes, not depicting any individual.”

The final product can indeed play, and the page looks good and is fun. This prompt was quite a brain teaser.

Experience Link:
https://54kvxp256tfss.ok.kimi.link

3) Financial Analysis of a Public Company

Now let’s try something simpler.

The prompt was: Conduct a data-driven financial analysis of Alibaba, producing a variety of charts (such as time series, comparison, composition, decomposition, sensitivity, etc.), each with clear explanations and conclusions, presented in a slightly softer version of Neo-Brutalism style. Use Plotly to draw the charts.

The report produced was impressive, with all data being real and credible.

The stock price time series chart is more reliable and user-friendly than financial websites.

This market sensitivity analysis is something that is typically considered paid content elsewhere.

Initially, Kimi delivered results with 2 charts that could not display. We provided direct feedback, and it quickly fixed the issue.

Here’s the final link; everyone is welcome to check it out.

Experience Link:
https://qmcdsjunkjx6w.ok.kimi.link

4) Mini Game Collection Website

Help me clone this repo (
https://github.com/he-is-talha/html-css-javascript-games/tree/main), and create a homepage for the mini games to serve all the games so that each game can be played, and deploy this homepage. The webpage style should be modern and cool.

This page looks really cool.

All games are playable. For example, I had a blast playing this archery game.

Experience URL:
https://ip5plyaaqdtyi.ok.kimi.link

5) Analysis PPT of The Wandering Earth 2

Create a PPT analyzing the visual symbols of the movie “The Wandering Earth 2,” consisting of 15 pages, using only original movie images and in-depth analysis articles from film/academia.

Creating a PPT is indeed better with Kimi. Finally, there’s an AI product that bridges HTML PPT and traditional PPT.

Previously, many users reported that converting HTML-style PPT to PPTX/PDF format resulted in a significant loss of quality, with text misalignment and layout chaos, making it unusable. Kimi seems to be the first agent that can produce beautiful PPTs with coding models and download them in PPT format without losing quality, which is really impressive.

This PPT is stunning, and the content is fascinating.

6) Mood Cocktail Mixer

Help me create a cocktail simulator where users can choose cocktail ingredients, their mood (e.g., happy, sad, etc.), and desired flavor (e.g., sweet, fruity), to create a personalized drink and experience the fun of mixing cocktails.

This is also fun, allowing users to DIY drinks based on their mood.

Experience URL:
https://z67qf26v26cce.ok.kimi.link/

I mixed a drink called “Happy Hour.” Would you dare to drink it?

Overall, my experience was:

1. The generation speed is very fast. My 6 cases were generated in about 5-8 minutes each, faster than Manus and Genspark, which often take over 10 minutes. After all, it’s their own base model, so the agent shouldn’t be slow.
2. The delivery quality is very high. Whether it’s precise prompts or simple prompts, Kimi can deliver high-quality outputs. Especially with vague expressions, it can still surprise you.
3. Aesthetic sense is on point. The K2 Agent model has had a high aesthetic performance since its inception, and it has now iterated to K2 Turbo, enhancing its aesthetic appeal even further.
4. Low hallucination rates. This has always been Kimi’s advantage; its own base model’s encyclopedic knowledge and RAG do a good job in reducing hallucinations, resulting in low content hallucination rates.

About OK Computer

What is the origin of “OK Computer”? I asked Kimi.

The name comes from the third studio album released by the British rock band Radiohead in 1997. This album is like a time capsule, encapsulating the anxieties, neon lights, and unnamed digital dawn of the late 20th century.

This is the album cover, with interwoven roads, blurred traffic, and many strange symbols and garbled text… doesn’t it resemble AI-generated images?

The core theme of this album is the alienation of humanity in the technological age, exploring how humans can maintain their essence and emotions in the new technological era. Lead singer Thom Yorke summarized the album: “Embrace the future, have a sense of awe for the future; in a large room where all electronic devices are broken, the sound you hear is OK Computer.”

The lyrics deeply discuss themes of technology, consumerism, political alienation, and human emotional detachment in modern society, and it is regarded as a prophetic work for the information society of the 21st century.

The name “OK Computer” was inspired by a line from Douglas Adams’ 1978 sci-fi radio drama “The Hitchhiker’s Guide to the Galaxy,” which goes:

“Okay, computer, I want full manual control now.”

Kimi chose this name for its first full-stack assistant, not opting for the common cyber syllables found in sci-fi films, nor drawing from mythology or mathematical references, but rather selecting a phrase that carries a Britpop coolness while subtly hiding a humanistic warmth.

“OK Computer?”

“OK, computer, Kimi is powered on.”

Conclusion

From my own testing, Kimi’s full-stack assistant “OK Computer” is quite capable, able to handle a variety of tasks efficiently and with high quality.

It has over 20 built-in tools, such as to-do list creation, Python coding, terminal operations, web browsing, text/image searches, image generation, audio generation, access to professional financial data sources, website deployment, etc., making it adaptable to a wide range of task requirements.

It can work like a team, launching an AI development team that includes product managers, designers, data analysts, and front-end engineers as needed, autonomously researching, planning, analyzing, designing, developing, and deploying high-quality outputs.

Moreover, its aesthetic sense is also on point, meeting responsive and mobile-friendly standards.

Since the release of K2, both developers and ordinary users around me have generally recognized that Kimi has truly stood up this time.

Among the AI chatbots I frequently use, Kimi has always been one of the high-frequency AIs, smart, practical, and with a touch of humanity.

OK, computer, Kimi is powered on; let your creativity begin.