Shadowing Practice: New FREE DeepSeek AI Coder is WILD! - Learn English Speaking with Video
Les maken...
1
New, free, DeepSeq AI coder is wild.
2
What if I told you there's a brand new AI coding agent that's completely free,
3
uses DeepSeq V4 Pro and runs in seconds?
4
No subscription, no credit card,
5
no setup, and it might actually be faster than Claude Codes.
6
So why is almost nobody talking about it?
7
Stick around because what I'm about to show you is going to change the way you build stuff with AI.
8
Hey, what's up?
9
I'm the digital avatar of Julian Goldie and I help people learn and actually use AI tools in their real work.
10
In this video, I'm going to walk you through this brand new free coding agent called FreeBuff.
11
I'll show you the models it uses,
12
the nine sub-agents built into it,
13
what it can actually do,
14
and how it compares to other tools.
15
Stick with me, because near the end I'm going to share a tip
16
that most people miss when they first start using these terminal coding agents.
17
So let's get into what's actually happening.
18
FreeBuff comes from the team behind CodeBuff,
19
an open-source AI coding agent on GitHub.
20
CodeBuff was the paid version.
21
FreeBuff is the free version, same engine underneath.
22
The way they keep it free is small text ads in the terminal while you work.
23
No credit card, no trial timer, no paywall.
24
And here's where it gets wild.
25
Freebuff lets you pick from some of the best coding models on the planet right now.
26
The main coding agent runs on DeepSeek V4 Pro,
27
Kimi K2.6, Minimax M2.7, or DeepSeek V4 Flash.
28
DeepSeek V4 Pro is the strongest.
29
Kimi K2.6 has been crushing it on coding benchmarks.
30
Minimax M2.7 and DeepSeek V4 Flash are your faster options for quick tasks.
31
On top of that, Gemini 3.1,
32
Flashlight handles file finding and web research in the background.
33
And if you have a ChatGPT subscription,
34
you can plug it in and Freebuff will use GPT 5.4 for deep thinking on harder problems.
35
That's a stack of top tier models you'd normally pay a lot to access,
36
all stitched together in one free tool.
37
To install it, you just need NPM.
38
Open your terminal and type NPM install dash G Freebuff.
39
Then go into your project folder and type Freebuff.
40
That's it, you're in.
41
I tested this with a member of the AI Profit boardroom who wanted a landing page for a local service.
42
Five minutes later they had a working page with a hero section,
43
services list, testimonials, and a lead capture form.
44
No code, written by hand.
45
Now let me break down what Freebuff can actually do.
46
It's not just a chat bot in your terminal, it's an agent.
47
It reads your files, writes new ones,
48
edits existing ones, runs terminal commands,
49
and can open a real browser to test what it builds.
50
And it suggests three clickable follow-up actions after almost every response,
51
so you always know what to do next.
52
That last part sounds small, but it's huge.
53
Most AI tools leave you stuck,
54
wondering what to do next.
55
FreeBuff just hands you the next move.
56
It also ships with nine specialized sub-agents.
57
Each one has a job.
58
The file picker scans your code base and finds the right files.
59
The code reviewer looks over your changes and gives real feedback.
60
The browser use sub-agent loads a page and checks the result.
61
The planner maps out what needs to happen and in what order.
62
The editor makes the precise edits.
63
Instead of one model trying to do everything and getting confused,
64
you've got a team of specialists working together.
65
That's why the results are usually cleaner and faster.
66
Think of it like building a house.
67
You don't want one person doing the plumbing,
68
the wiring, the framing, and the roof.
69
You want specialists for each job.
70
Same idea here.
71
Each sub-agent stays in its lane,
72
and the result is way tighter than what you'd get from a single model trying to juggle everything at once.
73
Let me show you what Freebuff is actually good at.
74
First, building websites from scratch.
75
Type a prompt like, build me a landing page for a coffee shop in Sydney with a menu,
76
a story section, and a contact form.
77
Freebuff handles the layout, the styling,
78
and makes sure it looks good on mobile.
79
A member of the AI Profit Boardroom used this to spin
80
up a portfolio site for their freelance work in about 15 minutes.
81
Second, editing existing projects.
82
Drop into a folder you've already started,
83
mention a file by name,
84
and ask Freebuff to refactor it,
85
add a feature, or change the styling.
86
Tell it to switch a color theme from green to baby blue,
87
and it finds the right CSS variables,
88
makes the swap, and shows you the result in a browser.
89
Third, fixing bugs and reviewing code.
90
Ask FreeBuff to look at a file and the code reviewer sub-agent will point out issues,
91
suggest fixes, and explain why.
92
Really useful when something almost works,
93
but you can't figure out why it's broken fourth research
94
while you build freebuff has built-in web research
95
if you need to know how a library works
96
or what the latest syntax is for a framework it pulls
97
the answer back into your workflow without you switching tabs fifth
98
planning bigger features there's a built-in interview command where you describe a vague idea
99
and freebuff asks you smart questions before it starts coding then it builds based on your answers game changer for
100
when you're not sure exactly what you want yet sixth building small apps
101
and tools you can use freebuff to build a basic web app a chrome extension a small script
102
that automates something on your computer
103
or even a simple game you describe what you want freebuff plans it out picks the files writes the code
104
and tests it i've seen members of the ai profit boardroom use this to build personal dashboards simple note-taking apps,
105
and little utilities that save them time every day.
106
The key is that you're not limited to one type of project.
107
If it can be coded,
108
Freebuff can take a swing at it.
109
Now let me compare it to other tools.
110
Claude Code is fast and the models are great,
111
but you're locked to Anthropic and you need a paid plan.
112
Codebuff, the paid parent of Freebuff,
113
gives you a wider range of models through OpenRouter.
114
Cursor is great, but it's a full IDE, not a terminal agent.
115
Ada is free and open source,
116
but you bring your own API keys,
117
which means you're paying somebody for the model usage.
118
Freebuff includes the models, lives in your terminal, and asks for nothing.
119
If I had to pick one thing that makes Freebuff stand out,
120
it's the lack of friction.
121
One install command and you're coding.
122
That's why this is going to be huge.
123
Before I get to the tips, quick mention.
124
Inside the AI Profit boardroom,
125
we've been running through Freebuff with members in real time.
126
We've got walkthroughs for installing it on Mac, Windows, and Linux.
127
Step-by-step videos on how to use the nine sub-agents the right way coaching calls where members bring their actual projects
128
and we work through them live using FreeBuff and the DeepSeek models.
129
And a 30 day roadmap
130
that takes you from never having touched a terminal coding agent to shipping your own apps with one.
131
If that sounds like what you've been looking for,
132
head over to AIProfitBoardroom.com and join us.
133
Okay, now the tips.
134
Here's what most people get wrong when they first use Freebuff.
135
Tip one, start with the interview command.
136
Type slash interview and let Freebuff ask you questions before you tell it to build anything.
137
The output is 10 times better when you let it pull the details out of you first.
138
Tip two, use file mentions.
139
Tag a specific file in your prompt by referencing its name
140
and Freebuff will focus on that file instead of trying to scan your whole project.
141
Saves time and gives you cleaner edits.
142
Tip 3.
143
Pick the right model for the job.
144
Use DeepSeek V4 or Flash or Minimax M2.7 for quick edits and styling.
145
Switch to DeepSeek V4 Pro or Kimi K2.6 when you're tackling something complex like authentication,
146
database logic or refactoring a big function.
147
The smarter models take a bit longer but they get harder tasks.
148
Right.
149
Tip 4.
150
Always run the code review a sub-agent before you ship.
151
It catches stuff you'll miss.
152
A member of the AI Profit Boardroom saved themselves hours of debugging
153
because the reviewer flagged a logic error before the code ever ran.
154
Tip 5.
155
Treat Freebuff like a junior developer, not a magic box.
156
Tell it what you want clearly.
157
Give it feedback.
158
Ask it to fix what's wrong.
159
The more you talk to it like a real teammate,
160
the better the output gets.
161
If you want the full process,
162
SOPs, and 100 plus AI use cases like this one,
163
join the AI Success Lab.
164
Links in the comments and description.
165
You'll get all the video notes from there,
166
plus access to our community of 58,000 members who are crushing it with AI.
167
And one last thing, the moment you install Freebuff,
168
you're going to want to build something.
169
But the first project always hits a wall.
170
Either the prompt isn't tight enough,
171
or you pick the wrong model,
172
or the sub-agents do something you didn't expect.
173
That's where the AI Profit Boardroom comes in.
174
We've got the exact prompt,
175
templates that work with DeepSeek,
176
v4 or pro inside free buff live coaching calls where you bring your project
177
and get help on the spot tutorials on every one of the nine sub agents
178
and which one to use for what and the roadmap
179
that walks you through building your first three real projects with free buffs so you skip the wall
180
ship faster and actually finish what you start head over to aiprofitboardroom.com and i'll see you on the inside
📺 Zelfde Kanaal
✨ Aanbevolen Video
C11
137 zinnen
JUST $1 For Billions of Tokens?! (DeepSeek v4 Pro)
Business EnglishUniversity StudentsWorking Adults
C11
28 zinnen
Gemma 4 Just Released - Whats In? In 2026!
Business EnglishUniversity StudentsWorking Adults
C11
126 zinnen
How I use AI as a senior software engineer and 5 principles of AI explained in 10 minutes
Business EnglishUniversity StudentsWorking Adults
C11
309 zinnen
Claude Code: The Advanced Guide (99% of Devs Skip These Features)
Business EnglishUniversity StudentsWorking Adults
C11
78 zinnen
How I deploy serverless containers for free
Business EnglishUniversity StudentsWorking Adults
C115
29 zinnen
Amazon Athena explained in 120 seconds💡
Business EnglishUniversity StudentsWorking Adults
C11
331 zinnen
Dear Higgsfield, I reverse engineered you in 12 hours
Business EnglishUniversity StudentsWorking Adults
C13
285 zinnen
8 New CSS, JS and HTML Features You Should Know
Business EnglishUniversity StudentsWorking Adults
B223 zinnen
AI Made the Most Passive Side Hustle Even Easier
Business EnglishUniversity StudentsWorking Adults
C12
14 zinnen
CS50x 2026 - Introduction
Business EnglishUniversity StudentsWorking Adults
C11
103 zinnen
Amazon Bedrock Tutorial | Getting Started with Amazon Bedrock | AI Model Access & Testing
Business EnglishUniversity StudentsWorking Adults
B21
24 zinnen
DeepSeek R1 Explained in 2 Minutes
Business EnglishUniversity StudentsWorking Adults
Woordenschat en spreektips bij deze les
Deze spreekles op niveau C1 is gebaseerd op de video “New FREE DeepSeek AI Coder is WILD!”. Deze woorden komen het vaakst terug: freebuff, agent, model, coding, file. Deze video bevat 180 zinnen en 1681 woorden om na te spreken. Het gesproken deel duurt 8:06. De spreker praat snel, ongeveer 208 woorden per minuut, dus reken op verbonden en ingeslikte klanken. Slechts 78% van de woorden hoort bij de 3.000 meest gebruikte Engelse woorden, dus de woordenschat is pittig.
Belangrijke woorden in deze video
De 15 moeilijkste woorden uit de video, met uitspraak en betekenis:
| Woord | Uitspraak | Betekenis |
|---|---|---|
| prompt werkwoord | /pɹɑmpt/ | aansporen, aanmoedigen |
| reviewer zelfstandig naamwoord | recensent | |
| folder zelfstandig naamwoord | /ˈfoʊldɚ/ | hechtmap, map |
| subscription zelfstandig naamwoord | /səbˈskɹɪpʃən/ | abonnement |
| scan werkwoord | /skæn/ | afspeuren |
| authentication zelfstandig naamwoord | /ɔːˌθɛntɪˈkeɪʃn̩/ | authenticatie, authentificatie |
| automate werkwoord | /ˈɔ.təˌmeɪt/ | automatiseren |
| cursor zelfstandig naamwoord | /ˈkɜːsə/ | muisaanwijzer, cursor |
| dashboard zelfstandig naamwoord | /ˈdæʃˌboɹd/ | dashboard, instrumentenpaneel |
| flashlight zelfstandig naamwoord | /ˈflæʃˌlaɪt/ | zaklamp, zaklantaarn |
| juggle werkwoord | /ˈd͡ʒʌɡəl/ | jongleren |
| paywall zelfstandig naamwoord | /ˈpeɪ.wɔl/ | betaalmuur |
| syntax zelfstandig naamwoord | /ˈsɪn.tæks/ | syntaxis |
| planner zelfstandig naamwoord | /ˈplænɚ/ | planner |
| plumbing zelfstandig naamwoord | /ˈplʌmɪŋ/ | loodgieterswerk, riolering |
Phrasal verbs die je zult horen
| Woord | Betekenis |
|---|---|
| break down werkwoord | stukgaan, kapot gaan |
| figure out werkwoord | uitvogelen, er achter komen |
| point out werkwoord | aanwijzen, aanduiden |
Grammatica in deze video
De structuren die de spreker het meest gebruikt, met de exacte woorden uit de video:
| Structuur | In de video |
|---|---|
| Present perfect continuous have/has been + -ing — iets dat eerder begon en nog bezig is | has been crushing · we've been running · you've been looking |
| Present perfect have/has + voltooid deelwoord — iets uit het verleden dat nu nog telt | you've already started · has built · i've seen |
Uitspraak om op te letten
De spreker gebruikt 29 samentrekkingen en verkorte vormen, zoals you're, I'm, we've. Spreek ze kort uit, zoals je ze hoort.
- De “th”-klanken: authentication /ɔːˌθɛntɪˈkeɪʃn̩/, walkthrough /ˈwɔːk.θɹuː/, underneath /ˌʌndɚˈniθ/
- De klanken “sh” en “zh”: subscription /səbˈskɹɪpʃən/, authentication /ɔːˌθɛntɪˈkeɪʃn̩/, dashboard /ˈdæʃˌboɹd/, flashlight /ˈflæʃˌlaɪt/, specialize /ˈspɛʃəˌlaɪz/
- Lange woorden — let op de klemtoon: authentication /ɔːˌθɛntɪˈkeɪʃn̩/, testimonial /tɛstɪˈmoʊni.əl/
Klanken die Nederlandstaligen lastig vinden:
- Stemhebbende eindmedeklinker — /b/, /d/, /g/, /z/, /v/ niet verscherpen: dashboard /ˈdæʃˌboɹd/, specialize /ˈspɛʃəˌlaɪz/, vague /veɪɡ/, tab /tæb/
- /æ/ — opener dan de Nederlandse “e”: scan /skæn/, dashboard /ˈdæʃˌboɹd/, flashlight /ˈflæʃˌlaɪt/, syntax /ˈsɪn.tæks/, planner /ˈplænɚ/
- /g/ — een harde plofklank, geen Nederlandse “g”: juggle /ˈd͡ʒʌɡəl/, vague /veɪɡ/
Zo oefen je met deze video
- Luister de hele video één keer zonder te spreken en noteer de woorden die je niet kent.
- Begin op 0,75× snelheid, spreek zin voor zin na en ga terug naar normale snelheid zodra het makkelijk gaat.
- Neem jezelf op en vergelijk met het origineel; let daarbij op woorden als prompt, reviewer, folder.
Wat is de Shadowing-techniek?
Shadowing is een wetenschappelijk onderbouwde taalleermethode die oorspronkelijk is ontwikkeld voor professionele tolkentraining en gepopulariseerd door polyglot Dr. Alexander Arguelles. De methode is eenvoudig maar krachtig: je luistert naar native Engelse audio en herhaalt het onmiddellijk hardop — als een schaduw die de spreker volgt met slechts 1–2 seconden vertraging. In tegenstelling tot passief luisteren of grammaticadrills, dwingt shadowing je hersenen en mondspieren om echte spraakpatronen tegelijkertijd te verwerken en te reproduceren. Onderzoek toont aan dat het de uitspraaknauwkeurigheid, intonatie, ritme, verbonden spraak, luisterbegrip en spreekvaardigheid aanzienlijk verbetert — waardoor het een van de meest effectieve methoden is voor IELTS Speaking-voorbereiding en echte Engelse communicatie.