I spent three whole nights making this' trading time machine 'myself, I never wrote a single line of code throughout the entire process, relying solely on Vibe coding: It can pause like playing a movie, accelerate forward and backward, drag the progress bar Even clicking on buying and selling points, any point on the profit and loss curve of funds, Instantly jump to any point in the history of your account, Visualize the buying and selling records at that time Information such as order placement location, position status, net exposure, etc, It can support any time level from daily to 1-minute. Partial visualization details include: Transparency and size visualization are used to distinguish quantities in buying and selling transactions Visualize the degree of light/heavy positions by using energy bars to visualize the net opening of positions (Proportion of position size to total funds) ... For example, you can click on May 19th, 2020, I can directly see on the candlestick what I did that day, how I did it, and why I made a loss, The memories and feelings of that time were suddenly awakened, Even the interface simulates the feeling of a real exchange (since it's called a time machine). The understanding of "backtesting" and "retrospective" in traditional stereotypes is completely different. I started learning Vibe coding from scratch and took many detours. The general process is as follows: First, roughly describe the conceptual requirements of the tool, and then throw it to Google AI Studio, Antigravity, Cursor, Even the Codex web version I found that the front-end page provided by Google AI Studio was the most reliable, so I continued on it, It was found that after excessive contextual clutter, the web version would have a dead loop: It starts by only writing code without submitting it, no matter how much it curses, it's useless, And after all, it is a web version, and the efficiency of use decreases linearly with the increase of project complexity. So I packaged the code and threw it to the Antigravity Windows version, It was found that Antigravity is currently only a preview version, After the free quota is used up, it is temporarily not supported to set up self paid Ai Agent API By now, we have completed approximately half of the current progress. So I packaged the code again and threw it to the cursor, Discovered that there is regional shielding in the cursor, and further studied how to solve it with a ladder .. It was found that using GFWList mode for the ladder is not feasible, and a global or regional whitelist must be used. Finally silky smooth. During the period, I felt: ChatGPT 5.1 is suitable for CTO, Gemini is suitable for front-end aesthetics and execution, Opus 4.5 is suitable for deep thinking on business logic problems, especially for data intensive requirements like mine. The current version's usage effect, such as ⬇️ As shown in the video The video was recorded spontaneously with voice explanations, but it was relatively rough, Even a bit incoherent, but with added subtitles for acceleration. You can roughly feel what concept tool this is through the video, It is a trading tool product form that most people have never seen in Web2 or Web3, After I reach a certain level of improvement, I will consider deploying it on a public website for everyone to experience and use. By the way, to summarize, during the week when I publicly tested API keys, Dori, Some tools spontaneously created by Twitter users: @0x0funky ： https://((((((x.com))))))/0x0funky/status/1993320956912451804 The first Bodhisattva, at the textbook level, shared the process of using my actual historical data for Vibe coding step by step, providing guidance to n other Twitter users, including myself, on trading analysis Vibe coding. The code is open source and deployed by @ okokgrok to a public website: https://(((((x.com)))/okokgrok/status/199347422396072052 @WeWill_Rocky： https://((((((x.com))))))/WeWill_Rocky/status/1994428045080592453 Visualization of current positions and pending orders,+ai analysis and summary, continuous product improvement and deployment on public websites. @Kuttielocs1 ： https://((((((x.com))))))/Kuttielocs1/status/1994037491549249919 The historical data dashboard further developed based on @ 0x0fungy's open-source code has been deployed to a public website. @porounclemao ： https://((((((x.com))))))/porounclemao/status/1993604029130400054 Trader Play Method Analyzer ", based on @ 0x0fungy's achievements, self refactored and+ai, detailed the thought process, and open sourced the code @semiokshen ：https://((((((x.com))))))/semiokshen/status/1994426107601547305 It is also a trading time machine concept, focusing on historical pending order analysis. Open source code. I'm not sure if there were any omissions. For those who have already produced results based on real historical data, if needed, I can continue to provide dedicated API keys that have been used for a longer period of time. Those who have relevant needs can bring their production results DM Twitter users can create so many interesting things purely based on their spontaneous enthusiasm for building, and they all openly and selflessly share them for free. This industry's trading field, especially full-time builders of To C products, should work harder. Finally, I can only say: Open Intelligence is amazing