In partnership with
Would you use a powerful AI for free if it showed ads?
WTF is ad-supported AI? Every prompt still creates a token bill. FreeBuff pays that bill, shows a small text ad between agent turns and uses the ad revenue to cover your usage.
You pay with attention instead of a monthly subscription.
DeepSeek added V4 Pro to its API on August 13.
A few hours later, FreeBuff founder James Grugett posted something stranger:
We are offering DeepSeek V4 Pro 100% free. It is free forever because we put ads in our harness.
FreeBuff is a separate company. DeepSeek still charges it for the tokens.
The ad works only if the average token bill is cheap enough to cover.
I am Alex. Welcome to ShortCu8 by Innov8.
Lets Dive Deep🐰🐰.
⭐Today's Shortcut
Try DeepSeek V4 Pro through FreeBuff before buying another AI plan.
Use it first for public or non-sensitive work. Free access comes with ads, account limits and a privacy trade.
Try it in your browser
Open FreeBuff and choose Chat.
Sign in, open the model picker and select DeepSeek V4 Pro if it is available for your account and country.
If it does not appear, FreeBuff says unsupported countries may receive a limited mode with V4 Flash and daily session limits.
Start with a public document, a research question or a small idea. Do not make a private client repository your first experiment.
Use it inside a project
FreeBuff also has a terminal agent. Install it with:
npm install -g freebuff
Then open your project folder and run it:
cd your-project
freebuff
The core product needs no API key or credit card. Choose V4 Pro from the available models and give it a bounded first task:
Read this project, explain how it works and list the three highest-risk problems. Do not edit any files yet.
Who pays the bill?
DeepSeek publishes the price of every input and output token.
Under the V4 Pro rates taking effect on August 16, a request containing 10,000 uncached input tokens and 2,000 output tokens would cost roughly:
1.1 US cents during off-peak hours
2.1 US cents during peak hours
That is one exchange. A long agent job may create many requests and cost far more.
FreeBuff needs ads, limits and model routing to cover an average bill measured in cents.
Capability is getting cheaper
DeepSeek is part of a wider price squeeze.
OpenAI prices GPT-5.6 Luna at $0.20 per million input tokens and $1.20 per million output tokens. The same 10,000-input and 2,000-output example costs about 0.44 cents.
GLM 5.2 costs $1.40 for input and $4.40 for output per million tokens. That example costs about 2.3 cents.
Kimi K3 and Qwen3.6 add another pressure: their weights are public, so more providers can host them and compete on price.
These models are not interchangeable. Price alone cannot tell us which one will finish a difficult job.
The useful signal is simpler. Models built for serious reasoning and agent work can now process an ordinary turn for fractions of a cent or a few cents.
What "free" costs
FreeBuff's privacy policy says prompts and messages may be analysed to personalize ads. Code pasted directly into a message can be part of that analysis.
It says uploaded files and connected repositories are not given to advertising providers. Your data may still pass through FreeBuff and the selected model provider.
FreeBuff also says it does not use prompt and project data for AI training unless the model or feature is clearly labelled for that use.
Use it for learning, experiments and work you are comfortable sending through those systems. Keep passwords, private keys and client secrets out of it.
Free AI is starting to borrow the economics of search.
The user sees a zero-dollar interface. An advertiser covers the average inference bill. Heavy use, private work and priority access move into paid plans.
The token meter is still running.
You just do not see the receipt.
Now go and test it 🌾🌾
The ShortList
🛠️Cool Tools of the Week:
X: The social media platform is expanding its open source codebase, letting users see if they’ve been "shadowbanned."
Runway: The video platform now connects Figma, Dropbox and Notion to its agent.
GPT-5.6 Sol: OpenAI has debuted an "ultrafast" version of its most powerful model, now running up to 14 times faster.
Perplexity Search as Code: The AI firm released optimizations to its deep research tool to improve performance while cutting cost per task by nearly 10%.
📩 Innathe Shortcu8 engane undarunnu 👇️?
We read every reply - just reply to this email and let us know how we can improve !
Appo adutha Shortcu8il kanaam bie…👋
If you read till here, you might find this interesting
#AD1
Stop making AI decisions in the dark.
Leadership is asking: are we getting value from AI? Which tools are worth the spend? Where are we exposed? Right now, most teams have no idea.
Harmonic Security Usage Explorer changes that.
You get a complete picture of how your organization uses AI, automatically categorized into custom tasks and use cases.
You’ll see the projects being worked on, who’s using what tools, where AI investments are driving value, and where employees are engaging in risky behavior.
CIOs can rationalize spending and cut wasted licenses. CISOs can pinpoint where risk exists and neutralize it. AI committees can show exactly how their efforts are paying off.
#AD2
Build a Holiday Creator Affiliate Program in 90 Days
Creators lock in holiday content calendars 90 days out, before brands figure out commissions. Waiting too long to launch an affiliate program means less runway to build demand and a missed shot at the best partnerships.
The 90-Day Holiday Sprint covers commissions, recruiting, and scaling a program at Day 30, 60, and 90.