start	end	text
0	4080	In this tutorial, we'll install OpenAI's Codex app on Windows.
4860	11200	Then, using CC Switch and AI HubMix, we'll make it run on almost any AI model, not just OpenAI's.
12720	15640	Start in your browser. Search for Codex Download.
15640	31600	Open the official Codex page on openai.com.
41980	43700	Click Download for Windows.
45640	56760	The installer itself is tiny, so it's done in seconds.
56760	62460	When it finishes, show it in File Explorer.
62460	70840	Right-click the installer and run it as administrator.
70840	77920	A setup window opens, and the real download begins.
77920	96520	This package is about 700 megabytes.
96520	100040	Fair warning, this part can be slow, usually 5-6 minutes.
100820	101700	So let's skip the wait.
109780	112480	And we're back, 96% and counting.
121720	124880	The moment it hits 100, everything happens on its own.
126520	145260	The installer closes, and Codex launches itself.
145260	146900	It lands on a sign-in screen.
146900	148240	And here's the key part.
148240	149240	Don't click anything.
149240	151060	We're never going to log in.
151060	154280	That's what CC Switch is for.
154280	155280	Codex is installed.
155280	157240	Now for the fun part, CC Switch.
158020	162320	CC Switch is a free, open-source app that can point codecs at third-party providers.
163120	164980	We'll use it to connect to AI HubMix,
165340	169900	one gateway that carries basically every AI model behind a single API key.
170680	175480	Today's example model, DeepSeq V4 Flash, the 0731 release.
177080	179920	Head to the CC Switch site and hit Download Free.
179920	186200	It takes us to the downloads page.
188280	189700	Pick your operating system.
190100	193180	We're on Windows 10, so we take the X64 installer.
196760	198980	Chrome flags the file as unverified.
199380	201200	That's just safe browsing being off.
201640	201960	Keep it.
207400	208540	12 megabytes.
208820	209380	Done instantly.
209380	216380	Open it to install.
216380	220380	Setup takes a few seconds, then CC switch is ready.
220380	222380	Open it up.
222380	225380	At the top, switch to the OpenAI section.
225380	227380	That's the one codex reads.
227380	231380	Then hit the plus to add a provider.
231380	234380	From the presets, choose AI HubMix.
234380	237380	Icon and endpoint fill in automatically.
239380	249560	Next we need an API key, so over to the AI HubMix website.
249560	256660	Search for iHubMix and open AIHubMix.com.
256660	265240	One gateway, infinite models, every major LLM behind one unified API.
265240	269360	Click Console.
269360	281340	It asks us to sign in Google account, one click.
281340	295140	Once we're in, open API keys in the sidebar and add a new key.
295140	304860	Name it DeepSeq V4 Flash 0731 so we know what it's for.
304860	308820	Notice the options here, quota and an expiry time.
308820	313620	We'll give this key an expiration date, one day.
313620	315700	If it ever leaks, it's already dead.
315700	317900	Cheap insurance.
317900	323480	Submit, and there's our key.
323480	329800	it before closing, it's only shown once. Back in CC Switch, paste the key into the AI HubMix provider.
331080	335960	Bonus feature, endpoint speed testing. Both endpoints answer fast, either one will do.
339000	345640	Now the default model. Click the download icon and CC Switch pulls the full AI HubMix model list,
345640	347640	everything they carry.
349640	355640	Filter for DeepSeq and select DeepSeq V4 Flash 0731.
357640	359640	Keep the upstream format on responses.
359640	361640	That's what Codex speaks natively.
362640	364640	Add it saved.
365640	366640	The new provider appears at the bottom.
366640	368640	Click it to enable it.
368640	370640	Now it's the active config for Codex.
370640	376640	Time to restart codecs properly.
376640	385640	Close the window, then quit it from the system tray, bottom right corner.
385640	394640	And launch it again.
394640	407000	While it starts, remember last time, it demanded a chat GPT sign in.
407000	409280	This time, straight to the welcome screen.
409280	410500	No login at all.
410500	416320	CC switch already wired codecs to AI HubMix.
416320	424220	Pick any field, it doesn't matter, and continue.
424220	425640	It offers a demo task.
425640	427640	Skip it.
427640	430220	Confirm.
430220	434980	Go to Chat GPT.
434980	435980	One last pop-up.
435980	438520	Pitching GPT 5.6.
438520	440140	Take Continue with Current Model.
440140	442020	So we stay on our custom setup.
442020	443400	Here's the main screen.
443400	444860	Bottom right of the input box.
444860	446620	The model says Custom.
446620	450460	That's our DeepSeq V4 Flash, riding through AI HubMix.
450460	451820	Time for a real test.
451820	464820	I'll ask for a short introduction of AI HubMix.
464820	467373	Questions in, send it.
468893	473173	The request flows through AI HubMix to DeepSeq and Codex gets to work.
473173	477613	It's an agent, so instead of guessing, it actually runs a shell command to look around
477613	478613	first.
482373	485473	21 seconds of work.
487753	488893	And here's the reply.
490953	491993	It's an honest one.
492373	495533	The model says it doesn't have reliable information about AI HubMix
495533	499293	and asks for a URL or some context instead of making things up.
500033	500633	Fine by us.
501033	502413	Next, we'll hand it the URL.
504333	507313	Picking up right where we left off, the model wanted a URL.
507753	508353	Let's give it one.
508353	515353	The answer, an AI model aggregator, and the URL is AIHubMix.com.
520213	521873	Sent. Now watch it work.
525373	530433	It announces the plan, fetch info from the site, and fires off Curl to download the homepage.
530433	537833	The first attempt hits a Windows quirk.
537833	541793	In PowerShell, curl is just an alias for InvokeWebRequest.
541793	549933	The model catches that and switches to the real curl.exe.
549933	560413	Still blocked, so it retries, this time pretending to be a browser with a Mozilla user agent.
560413	562413	This is what makes agents fun to watch.
562413	566413	It's debugging its own path to the answer.
570413	578413	And remember, this is DeepSeq v4 Flash, a fast, inexpensive model doing genuine multi-step work.
584413	589413	Once the page downloads, it strips the HTML down to readable text.
590413	599473	1 minute, 31 seconds of work.
601333	602553	And here's the payoff.
604153	606853	A real introduction, sourced from the live site.
607653	613513	One API, many models, and open AI-compatible endpoint, routing, developer tools, the works.
614213	614753	Final stop.
615133	616333	Let's see what all of this costs.
616733	618293	Back to the AI HubMix console.
618853	620333	In the sidebar, open logs.
620413	624093	Here is every request itemized.
624093	632593	Each row shows the model, our deep seek, v4, flash, 0731, tokens in and out, cash tokens,
632593	636193	latency around a second and a half, and the exact spend.
636193	639213	Most of these calls cost less than a tenth of a cent.
639213	643913	And that's the full setup, codecs on windows, CC switch in the middle, and one AI Hub Mix
643913	646733	key that unlocks any model you want to try.
646733	647413	Thanks for watching.