Kai

@hqmank • 14,270 subscribers

Kai Wang (AI Kai) AI Builder 🛠️ Practical AI Workflows & Coding Tips High-Value Promo Codes Updated Regularly DM for collab 📩

Shorts

Kimi K3 running inside the Codex app. Here is the 5-minute setup.

Kimi K3 running inside the Codex app. Here is the 5-minute setup.

164,867 次观看

Compiled the leaked Claude Code source and got it running locally. Then I spun up an Agent Team and had it analyze its own codebase.

Compiled the leaked Claude Code source and got it running locally. Then I spun up an Agent Team and had it analyze its own codebase.

254,035 次观看

ChatGPT macOS just updated. Ultra is gone from the effort slider. You have to go into settings to enable it, or set it up through Advanced mode. Codex did the same thing.

ChatGPT macOS just updated. Ultra is gone from the effort slider. You have to go into settings to enable it, or set it up through Advanced mode. Codex did the same thing.

88,635 次观看

PSA: Codex and ChatGPT Work are getting another usage reset. OpenAI says rate limits will reset twice in the next 24 hours for the GPT-5.6 Sol launch. If you have big tasks waiting, now is the time. Go full throttle.

PSA: Codex and ChatGPT Work are getting another usage reset. OpenAI says rate limits will reset twice in the next 24 hours for the GPT-5.6 Sol launch. If you have big tasks waiting, now is the time. Go full throttle.

64,674 次观看

You can use your ChatGPT subscription inside Cursor. Tibo publicly shared this method and supports CLIProxyAPI, so it's not some workaround that'll get patched. 1. Install CLIProxyAPI (proxies your ChatGPT sub as an OpenAI-compatible API) 2. Open Cursor > Settings > Models 3. Enter your CPA public URL and API key ⚠️ Two catches. You need Cursor Pro. Free tier won't let you add third-party API keys. Cursor rejects localhost URLs, so you need a public endpoint. Run cloudflared tunnel --url and you get one in seconds. Paste that into Cursor. Tested it. Works.

You can use your ChatGPT subscription inside Cursor. Tibo publicly shared this method and supports CLIProxyAPI, so it's not some workaround that'll get patched. 1. Install CLIProxyAPI (proxies your ChatGPT sub as an OpenAI-compatible API) 2. Open Cursor > Settings > Models 3. Enter your CPA public URL and API key ⚠️ Two catches. You need Cursor Pro. Free tier won't let you add third-party API keys. Cursor rejects localhost URLs, so you need a public endpoint. Run cloudflared tunnel --url and you get one in seconds. Paste that into Cursor. Tested it. Works.

29,358 次观看

Railway has a free cloud VM with Claude Code, Codex, Cursor CLI and other coding agents preinstalled. It's super easy to use. Just run one SSH command: ssh Each VM comes with 2 vCPU and 2GB RAM. You get 60 minutes to build, then 24 hours to claim it before the VM and its files are deleted. I tried it and recorded the setup, plus tests with several coding agents. You'll need GitHub login and an active Railway trial to claim the VM.

Railway has a free cloud VM with Claude Code, Codex, Cursor CLI and other coding agents preinstalled. It's super easy to use. Just run one SSH command: ssh Each VM comes with 2 vCPU and 2GB RAM. You get 60 minutes to build, then 24 hours to claim it before the VM and its files are deleted. I tried it and recorded the setup, plus tests with several coding agents. You'll need GitHub login and an active Railway trial to claim the VM.

11,976 次观看

Gave DeepSeek vision with one skill. npx -y skills add liustack/modlens Repo link 👇

Gave DeepSeek vision with one skill. npx -y skills add liustack/modlens Repo link 👇

37,561 次观看

Tried Pi Agent with DeepSeek V4 Flash. It’s really fast. DeepSeek lacks vision support, so I use a plugin to send visual tasks to Grok 4.5. Here’s a quick demo 👇

Tried Pi Agent with DeepSeek V4 Flash. It’s really fast. DeepSeek lacks vision support, so I use a plugin to send visual tasks to Grok 4.5. Here’s a quick demo 👇

26,299 次观看

MiniMax M3: Opus-level coding at DeepSeek pricing. On Terminal-Bench 2.1, it scores 66.0, only 0.1 behind Opus 4.7. I gave it a quick try on a few frontend tasks, and the output quality genuinely feels close to Opus 4.7.

MiniMax M3: Opus-level coding at DeepSeek pricing. On Terminal-Bench 2.1, it scores 66.0, only 0.1 behind Opus 4.7. I gave it a quick try on a few frontend tasks, and the output quality genuinely feels close to Opus 4.7.

25,730 次观看

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Everyone keeps talking about how good Kimi K3 is. So I gave it one of my favorite frontend tests: one prompt + one reference image. The task: rebuild a 3D globe dashboard. 🌎 The result was almost too good. It did not just build a working page. The texture, lighting, spacing, depth, glass panels, and overall layout were all remarkably close to the reference. Some parts felt nearly pixel-perfect. My honest first impression: for this kind of frontend work, Kimi K3 might be slightly better than Fable 5. All at Sonnet prices.

Everyone keeps talking about how good Kimi K3 is. So I gave it one of my favorite frontend tests: one prompt + one reference image. The task: rebuild a 3D globe dashboard. 🌎 The result was almost too good. It did not just build a working page. The texture, lighting, spacing, depth, glass panels, and overall layout were all remarkably close to the reference. Some parts felt nearly pixel-perfect. My honest first impression: for this kind of frontend work, Kimi K3 might be slightly better than Fable 5. All at Sonnet prices.

406,215 次观看 • 3 个月前

Every tip in Andrej Karpathy's post is gold. I use the HTML one a lot. When I'm learning something new or trying to understand complex code, I ask the agent to build me a visual HTML page. Here I was learning SEO and GEO. The agent turned the key ideas into an interactive page I could click through. I picked it up faster than I would have from reading docs.

Every tip in Andrej Karpathy's post is gold. I use the HTML one a lot. When I'm learning something new or trying to understand complex code, I ask the agent to build me a visual HTML page. Here I was learning SEO and GEO. The agent turned the key ideas into an interactive page I could click through. I picked it up faster than I would have from reading docs.

51,534 次观看 • 13 天前

Kimi K3's frontend work here is insane. Someone asked a Kimi K3 Max agent swarm to recreate macOS 27 as a web app. It looks so much like a real Mac that I kept forgetting the whole thing was running in a browser.

Kimi K3's frontend work here is insane. Someone asked a Kimi K3 Max agent swarm to recreate macOS 27 as a web app. It looks so much like a real Mac that I kept forgetting the whole thing was running in a browser.

271,192 次观看 • 3 个月前

Kimi K3 vs Fable 5. Same prompt and reference. 3D globe dashboard. Kimi left, Fable right. To me, Kimi's UI fidelity is a touch better. The globe looks better too, with richer textures and smoother motion. This task: ~$3 with Kimi, ~$15 with Fable. I'd pick Kimi k3 for frontend work.

Kimi K3 vs Fable 5. Same prompt and reference. 3D globe dashboard. Kimi left, Fable right. To me, Kimi's UI fidelity is a touch better. The globe looks better too, with richer textures and smoother motion. This task: ~$3 with Kimi, ~$15 with Fable. I'd pick Kimi k3 for frontend work.

226,853 次观看 • 3 个月前

GPT-5.6 Sol vs Fable 5. I put both through my hardest frontend test: a 3D globe dashboard. Same prompt. Same reference image. Sol on the left. Fable on the right. GPT-5.6 Sol is a major leap from 5.5. Fable 5 still wins on spacing, balance, and first-pass polish. For me, Sol comes close at half the cost.

GPT-5.6 Sol vs Fable 5. I put both through my hardest frontend test: a 3D globe dashboard. Same prompt. Same reference image. Sol on the left. Fable on the right. GPT-5.6 Sol is a major leap from 5.5. Fable 5 still wins on spacing, balance, and first-pass polish. For me, Sol comes close at half the cost.

199,872 次观看 • 3 个月前

Combined Jev with Playwright-controlled Chrome to build jev-browser, a general browser automation skill. Works across agents. Two demos in Antigravity CLI and Codex: finding related articles, then job openings on a site. Both ran fast. This could handle many more browser tasks.

Combined Jev with Playwright-controlled Chrome to build jev-browser, a general browser automation skill. Works across agents. Two demos in Antigravity CLI and Codex: finding related articles, then job openings on a site. Both ran fast. This could handle many more browser tasks.

45,354 次观看 • 28 天前

Fable 5 is back. So I gave it one of my favorite frontend tests: one prompt + one reference image. The task: rebuild a 3D globe dashboard.🌎 The result was almost too good. It did not just make a working page. The texture, lighting, spacing, depth, glass panels, and overall layout were all close to the reference. Some parts felt nearly pixel-level.

Fable 5 is back. So I gave it one of my favorite frontend tests: one prompt + one reference image. The task: rebuild a 3D globe dashboard.🌎 The result was almost too good. It did not just make a working page. The texture, lighting, spacing, depth, glass panels, and overall layout were all close to the reference. Some parts felt nearly pixel-level.

116,611 次观看 • 3 个月前

Qwen3.8 just dropped. I gave it one of my favorite frontend tests: One prompt + one reference image. The task: rebuild a 3D globe dashboard. 🌎 It built a working page. Layout, structure, all there. Texture looked decent. Lighting flat. Glass panels had no depth. Some elements didn't match the reference. On frontend, not close to top-tier models like Kimi K3.

Qwen3.8 just dropped. I gave it one of my favorite frontend tests: One prompt + one reference image. The task: rebuild a 3D globe dashboard. 🌎 It built a working page. Layout, structure, all there. Texture looked decent. Lighting flat. Glass panels had no depth. Some elements didn't match the reference. On frontend, not close to top-tier models like Kimi K3.

29,292 次观看 • 3 个月前

Did GPT-5.6 Sol beat Fable 5 for frontend work? GPT-5.5 got roasted for its frontend. Had to test the new one. Same challenge: one prompt + one reference image → Three.js 3D globe dashboard. Verdict: Fable 5 still edges it out. But at half the cost, GPT-5.6 Sol is a serious option now.

Did GPT-5.6 Sol beat Fable 5 for frontend work? GPT-5.5 got roasted for its frontend. Had to test the new one. Same challenge: one prompt + one reference image → Three.js 3D globe dashboard. Verdict: Fable 5 still edges it out. But at half the cost, GPT-5.6 Sol is a serious option now.

21,502 次观看 • 3 个月前

I've been trying a tool called treg. It handles one of the annoying parts of working with agents: finding the right API for a task. You describe the job. treg searches its catalog, shows the available providers and prices, then lets your agent call the selected endpoint with one token. I tested it with a question I would ask in real work: "What AI coding topics are trending on X right now?" The agent searched for an X post API, compared the options, picked one, pulled the posts, and summarized the topics. I like the idea behind it: "Ask for the task, not the tool." treg is OpenRouter for agent tools. - 2,630 endpoints across 47 providers - pay per call at provider rates, no subscription or markup - open source

I've been trying a tool called treg. It handles one of the annoying parts of working with agents: finding the right API for a task. You describe the job. treg searches its catalog, shows the available providers and prices, then lets your agent call the selected endpoint with one token. I tested it with a question I would ask in real work: "What AI coding topics are trending on X right now?" The agent searched for an X post API, compared the options, picked one, pulled the posts, and summarized the topics. I like the idea behind it: "Ask for the task, not the tool." treg is OpenRouter for agent tools. - 2,630 endpoints across 47 providers - pay per call at provider rates, no subscription or markup - open source

13,316 次观看 • 1 个月前

Kimi K3 vs GPT-5.6 Sol. I reused the 3D globe test from Kimi vs Fable 5. Same prompt. Same reference image. Kimi K3 left. GPT-5.6 Sol right. Kimi edges out Sol on UI fidelity and globe realism. At current API prices, Kimi is about 50% cheaper. For frontend work, it's my pick.

Kimi K3 vs GPT-5.6 Sol. I reused the 3D globe test from Kimi vs Fable 5. Same prompt. Same reference image. Kimi K3 left. GPT-5.6 Sol right. Kimi edges out Sol on UI fidelity and globe realism. At current API prices, Kimi is about 50% cheaper. For frontend work, it's my pick.

15,769 次观看 • 2 个月前

没有更多内容可加载