Mokoena

[Mokoena @wickusluckystar]

@Catherine_ki А у меня даже паспорт не спрашивали.

[Chioma Amadi @Chioma__Amadi]

If you’re hunting for a remote job, there are some recruitment agencies and talent networks that can connect you to employers who are already hiring.

All you have to do is create a talent profile, get matched with opportunities, or apply for roles through their platforms.

Here are some of those agencies and how to get started:

[Mokoena @wickusluckystar]

@Rashid971x Хороший способ привлечения трафика😂

[Mokoena @wickusluckystar]

@Rashid971x BTC

[布丁 @lengs4189]

听说最近 ChatGPT 又开始弹手机验证了？

不用担心，也不用去闲鱼找高价接码。我这个网站收集了 75 个接码平台，很多闲鱼卖家都是从这里进货，几分钱的成本卖几块钱。

有需要的收藏一下，少花点冤枉钱。

H5 端：https://t.co/LjBwtO3DyP
PC 端：https://t.co/xekTE2PPbX

[0xYAS @calbuldelis69]

A free terminal coding agent just launched with 50M tokens a week. Plus a couple other free options I haven't posted.

1. Directionally Local (dl)
• Open source terminal coding agent
• GPT-6 Luna + DeepSeek V4.1 Flash
• Up to 50M tokens per week
• No card, GitHub sign-in
• Newer GitHub accounts start on DeepSeek only (5M/week)
• Resets every Monday
• Invite friends for extra tokens
• Works on Mac, Linux, Windows WSL
• Install: curl -fsSL https://dl{dot}new | sh

2. MiniMax H3 free site
• Daily free MiniMax H3 videos
• Free images
• Not a coding API

3. LLM7{dot}io
• DeepSeek V4 Flash and GLM 5.3 Flash free
• Rate limited free token

4. FreeLLMAPI
• Aggregator routing to free DeepSeek endpoints
• One key, automatic failover

Links in the comments.

I'll keep testing and sharing the ones actually worth using.

Disclaimer: Third-party services. Offers, models, limits, and availability can change. Never send passwords, API keys, seed phrases, private keys, or sensitive data through free/public APIs.

[Moeez @MATKA_r]

50M free AI tokens every week for coding 😱

And it is completely open source

Directionally Local just launched dl a free coding agent that runs on your own machine.

Here's what you get:

→ 50M free weekly tokens on GPT-6 Luna and DeepSeek V4.1 Flash

→ No credit card required just GitHub sign-in

→ Resume Claude Code sessions with
"dl --from-claude-code"

→ Works on Mac, Linux, and Windows (via WSL)

→ Open source with no vendor lock-in

Get started with one command:

"curl -fsSL https://t.co/S6UF97ClYH | sh"

→ https://t.co/S6UF97ClYH

[Mokoena @wickusluckystar]

@wheregoing123 What is this?

[Mokoena @wickusluckystar]

@oskwwwww Me too 😔

[Mokoena @wickusluckystar]

@wheregoing123 And then, when I was about 50% done with the project, I got banned.
Meanwhile, an entire month of actively messing around with random bullshit hadn't attracted anywhere near that kind of attention.

[Mokoena @wickusluckystar]

@wheregoing123 I think I finally figured out why I got banned. About five days before the ban, I was heavily using Claude Code to build an Airtable clone. It was some pretty intense development work, involving a lot of reverse engineering.

[Mokoena @wickusluckystar]

@make_claude @AnthropicAI контора пидорасов! @OpenAI тоже контора пидорасов, и они это даже не скрывают! Я конечно пару дней по грустил, но сейчас взял себя в руки и пользуюсь другими моделями. Да, сложно! Да, тяжело дается продолжение большого проекта. Ноя адаптируюсь к этой новой реальности

[make claude @make_claude]

ANTHROPIC 是如何对待俄罗斯用户的

我把俄语区 Claude 用户正在面对的问题整理到了一起。每个问题单独列出，这样更容易看清整体情况到底有多糟糕。

1. 先付费后验证。订阅费先扣了，验证身份的流程是付完钱才出现的。如果你用的是虚拟卡或者礼品订阅，退款根本没有途径。这是不是巧合，不好说。

2. 验证过程中会采集生物识别数据。按照法律他们可以保存这些数据 3 年。验证结束后访问权限被关闭，但你的数据他们已经拿到了。

3. 硬编码封锁。你的 machineID 和 device_id 会被记录。邮箱被加入黑名单。与你账户关联的 GitHub 仓库也会被封。开发者不仅失去了 Claude，还没办法把项目移交给别人，也不能从另一个账户继续工作。

4. 移民同样被封锁。你已经离开了那个国家，住在欧盟，但需要用俄罗斯 IP 登录某个服务。算法检测到这一点就把你封了。你所有的工作、项目和数据全部丢失。

5. 这不只是遵守法律。Google 和 OpenAI 面临同样的限制，但他们不会对用户进行全面搜查。他们会在付款之前提醒用户。Anthropic 选择了另一条路，看起来更像是公司政策而不是法律要求。

6. 这不是法律，是企业内部规定。如果一家面包店拒绝卖面包给爱斯基摩人，这家店会被关掉。按地区封锁和按国籍封锁个人用户是完全不同的事。一个美国人在俄罗斯境内可以正常使用 Claude，因为他的国籍。一个持有合法居留许可住在欧盟的俄罗斯人却不行。这已经不是地理问题了，这是基于国籍的封锁。

7. 最荒谬的是，你们试图把俄罗斯隔绝在你们的技术之外，但你们并没有把自己隔绝在俄罗斯之外。石油、天然气、木材还是同样的量在流通，只不过经过了壳公司。封锁的严厉程度被执行的灵活性所抵消。

一个每月付 20 美元用 Claude 工作的用户，不应该为某些人没有解决他们的纷争而负责。这些决定不是普通用户做出的。但恰恰是这个用户失去了钱、数据和工具。而这家公司连公开提出的直接问题都不回答。

Anthropic 没有回复过任何公开提问。

你遇到过这样的情况吗？

P.S. 收藏这条，给那些觉得地区封锁只影响 VPN 用户的人看。转发给那些还在说 "你不违反规则就好了" 的人。

[Arif AI @ArifAIHQ]

50 legendary websites that feel like the internet’s hidden toolbox 🧰

1. https://t.co/fNfYPfqOQL — free research papers
2. https://t.co/RLNx06Hapr — papers when unpaywall misses
3. https://t.co/M7RwZWAPmh — borrow books
4. https://t.co/JGLJ1vBMd4 — 70k+ free books
5. https://t.co/cJPoYHngwT — audit university courses free
6. https://t.co/yrpzRcVq4X — step-by-step math
7. https://t.co/Wv5dNqTyKU — background removal
8. https://t.co/AfwRqOXsE6 — object removal
9. https://t.co/YeoVMU3HZq — quick file tools
10. https://t.co/6Weq5y7fyH — PDF merge/split
11. https://t.co/PSQWFnBqHS — free Photoshop alternative
12. https://t.co/dVHCcpl8zq — image compression
13. https://t.co/SRCxOs4XW5 — fast diagrams
14. https://t.co/ZxmRK0LWeE — beautiful code screenshots
15. https://t.co/jYuShbP6wE — test regex
16. https://t.co/VOzFrBSHzi — clean JSON
17. https://t.co/isVS4wQPrE — terminal commands explained
18. https://t.co/gOkFgZk8Bq — software alternatives
19. https://t.co/vORyZ5C5Xb — breach check
20. https://t.co/0iez5I3gqn — temp email
21. https://t.co/whAuSUyU4c — self-destruct notes
22. https://t.co/bDEEnoZEJh — live global radio
23. https://t.co/PFgUKsi4hz — focus sound
24. https://t.co/xEb9mdypc1 — custom ambient
25. https://t.co/oKnODbmu1k — where to stream
26. https://t.co/Q31kxdRdCm — digital library
27. https://t.co/SBOSbqzpJd — old web pages
28. https://t.co/dxb0bCZ1v0 — site traffic
29. downtimerobot — uptime check
30. https://t.co/1TKu00WSYX — speed test
31. https://t.co/cUsbPNz6FA — compress images
32. https://t.co/M4975OPTy9 — file conversion
33. https://t.co/Hkm25zHUzx — compare text
34. https://t.co/Sz27b2GxfH — disposable inbox
35. https://t.co/sOnhiTIpkq — JS obfuscate
36. https://t.co/G3oVUURj63 — typing practice
37. https://t.co/sOL07zfQGQ — typing test
38. https://t.co/EyJ9dB4Pvt — spaced repetition
39. https://t.co/XIqIGL8XN3 — all-in-one notes
40. obsidian.md — local knowledge base
41. https://t.co/J4HQwNWk9x — bookmark manager
42. https://t.co/JCPE1azurg — task list
43. https://t.co/4NZqyL0VGf — boards
44. https://t.co/fJju8lecx7 — whiteboard
45. https://t.co/EjxXWG4Nes — design
46. https://t.co/kFr6p0jmw1 — quick graphics
47. https://t.co/Wv5dNqTyKU — again because it's that good
48. yt-dlp — video download (open source)
49. https://t.co/Q386r6mcbC — video compress
50. https://t.co/9pcBBcY0gy — local file transfer

Save this. You’ll definitely need some of these later. 🔖
Follow @ArifAIHQ for more useful websites, AI tools & tech resources.

[Peng @pengsonal]

AN UNDERRATED AI API GEM WITH FREE CREDITS 🤯

HiLinkUP is another option worth checking out if you use OpenRouter

sign up here: https://t.co/Z1e7YaDUb7

what you get:

• Free signup credits, no credit card required
• One OpenAI-compatible API key
• Access to Claude, DeepSeek, Qwen, Kimi, GLM and more
• Up to 70% off selected models
• Works with Cursor, Claude Code, Codex and more

plus, the 100% top-up bonus doubles your credits:

• Top up $10 → Get $20 in credits
• Top up $20 → Get $40 in credits
• Top up $50 → Get $100 in credits
• Top up $100 → Get $200 in credits
• Top up $500 → Get $1,000 in credits

that's double the credits. check the bonus rules for expiry dates and usage limits.

API base URL: https://t.co/aTrT81c6ga

worth checking out if you use AI APIs regularly

[ＺＥＦＩ @zefirium]

nvidia is hosting frontier AI MODELS FOR FREE

no card for the trial key:

> nvidia nim: deepseek v4.1 flash, glm 5.3 flash, mistral large 4, step 5 preview. openai-compatible base url: https://t.co/8CvetTVutj

> openrouter free right now: gemma 4 31b + 26b, flint, xing 4.0, north mini code, dots3-note, lfm 2.5: https://t.co/0BPng3WnQQ

> opencode zen $0: deepseek v4.1 flash, glm 5.3 flash, mimo v2.6 flash, flint 40b, beam: https://t.co/zDdZJXnFb5

> orcarouter free: deepseek v4.1 flash + glm 5.3 flash (also hy4 preview): https://t.co/Jjb0JSYBvN

> kimi k3.1 pro landed on nvidia's catalog nov 4 with a free endpoint badge, but api access is still closed. playground only: https://t.co/gt3Upzwy4w

the catch:
• nvidia trial terms: testing and evaluation only, not production
• credits and throttling apply. they don't publish a fixed rpm on every model card
• pollinations listed mistral large 4 and ling 3.1 flash yesterday - both paid_only there, don't treat as free
• viral posts listing kimi k3.1 pro as a free api are ahead of the catalog. keys won't authorize it yet

point any openai-compatible agent at the nvidia base url (or openrouter/opencode/orca) and swap the model id

[Brucky @danikkk_wqs]

it's f*cking painful how much money people waste running Claude Code on one model, when Anthropic just showed the 3 model setup that does it better for less

in their tests, an Opus advisor more than doubled Haiku's score, at 85% less than running Sonnet alone

one command:

claude --advisor opus --subagents haiku

who does what:

→ Sonnet 5.5 runs the session and writes the code

→ Haiku 5.5 digs through files and docs in parallel

→ Opus 5.5 sits on the side as an advisor for the calls that matter

→ you stop burning Opus on work that never needed it

add this one line to your CLAUDE.md:

"consult the advisor before locking a plan, after the same error twice, and before marking a task done"

Sonnet builds. Haiku digs. Opus judges

bookmark this. it'll be your daily driver by next week

[Igor 🪓 @Kwieser00]

Chineses mais uma vez dando mentoria de graça. Nunca ignore um mercado com um bilhão e quatrocentos milhões de pessoas.

[cabbit @cabbitar]

this is absolutely crazy

this thing removes the censorship from any AI model right in your browser. for free.

obliteratus is the open-source tool Elder Plinius used to make the uncensored DeepSeek V4.1 Flash.

how to use it (no code):

1. Open the Hugging Face Space
https://t.co/YnmIBaNcJ4

2. Or open the Colab notebook and hit Run all
https://t.co/BC6L6s3ujW

3. Pick a model, obliterate it, then test it

for devs:
obliteratus obliterate meta-llama/Llama-3.1-8B-Instruct --method advanced

and you’re responsible for what you do with the model

save it
and check out the article below: