I recently tried the new DeepSeek V4 Flash, and honestly, it completely blew me away.
Here’s what happened: I had been using OpenAI models when I suddenly received an email saying that cyber abuse had been detected on my account.
I’ve been building a returned-ticket monitoring tool for the Rolex Shanghai Masters. I submitted an appeal right away, but I was also concerned that certain parts of the project might have crossed a line.
So I temporarily switched setups and used OpenCode with DeepSeek V4 Flash to finish some of the remaining optimizations.
My biggest takeaway: this model is seriously capable.
First, the cost is almost unbelievably low.
I used roughly 150 million tokens in total. Since most of them were cache hits, the final bill came to less than RMB 5—under one US dollar. I genuinely had to double-check because I thought I’d read it wrong 😂
Second, it’s incredibly fast.
DeepSeek V4 Flash has very high throughput, and many responses appeared almost instantly. The experience felt remarkably fluid—the model could keep up with the natural pace of my thoughts without long pauses interrupting my workflow.
More importantly, it isn’t just cheap and fast. When handling real optimization work on an actual project, it understood the requirements well and consistently moved the work forward.
Cheap, fast, and genuinely capable.
If you enjoy building fun side projects—or simply want a coding model that is affordable and highly responsive—I strongly recommend giving DeepSeek V4 Flash a try.
An absolute godsend 🙏
@abhiijayVinayak I didn’t intentionally choose Bing over Google. I submitted the site to both GSC and Bing Webmaster Tools, but Bing indexed and ranked my new free tool pages much faster. Google traffic is still limited, possibly because the domain is relatively new.
My SaaS has been live for 4 months, and I finally saw SEO start to move.
I'm still new to SEO, so take this as a founder note, not advice.
Over the last few months, I tried a bunch of things:
- writing blog posts, though not very consistently
- building comparison / alternative pages
- trying programmatic SEO for long-tail pages
- improving titles, descriptions, internal links, multilingual pages, and GEO
I had no ad budget, no existing audience for this product, and no real domain authority.
The surprising part:
free tools worked fastest.
In May, I started shipping dozens of small audio/video/subtitle tools:
video converters, audio tools, YouTube subtitle downloaders, thumbnail downloaders, MP4 to MP3, MKV to MP4, subtitle extractors, compressors, reversers, speed changers, etc.
The BWT screenshot shows the result:
- 15.5k Bing clicks
- 530k impressions
- 2.92% CTR
- from May 26 to Jul 7
PostHog showed the same direction:
Bing search pageviews went from 167 in April to 12.6k in June.
Google is still tiny. Maybe sandbox, maybe new-site trust, maybe I just haven't earned enough authority yet.
My current mental model:
Blog posts explain ideas.
Comparison pages capture selection intent.
Programmatic SEO covers long-tail variations.
Free tools capture urgent tasks.
The tool pages are not "content" in the traditional blog sense, but they answer the exact thing someone is trying to do right now.
"I need to convert this file."
"I need to download this subtitle."
"I need to compress this video."
That might be why they worked before the brand had much authority.
Important caveat:
This probably works better for some SaaS products than others.
SubHero sits next to a lot of tiny media workflows, so free tools are a natural wedge. Not every SaaS has dozens of useful standalone tasks people already search for.
So the question may not be "should every SaaS build free tools?"
It might be:
can your product's workflow be broken into small, useful, searchable jobs?
Curious how other bootstrapped founders think about this:
if you had almost no budget and were starting distribution from zero, how would you try to get the first wave of organic traffic?
@i_mika_el Yes, some of it did convert. A few of the free tools export subtitles or other audio/video files, so I added a CTA at the end that leads users to try our transcription and translation products. That flow has brought in real signups, not just traffic.