@grminicci
iAccount based inUnited States!
About this account
- Account based in
- United States
- Connected via
- China Android App
! X says this location may be affected by a proxy or VPN.
Account-level information from X, not a live location or the device used for a specific post.
成长中的技术极客,硬件攻城狮
Japan
Joined September 2017
- Tweets7.6K
- Following302
- Followers9
- Likes169
hello retweeted
ChatGPT Plus 免费 1 个月 😳
领取链接:chatgpt.com/?promo_campaign=…
这次是美区节点,其他区没有试,请自行测试
chatgpt 好像又开始一个月Plus免费试用活动了
刚才打开一个free老号突然出现试用
我看了下是要挂日本节点
如果你有之前的老号也可以试试,不保证所有的号都有
如果没有出现,可以试试触发链接:
chatgpt.com/?promo_campaign=…
hello retweeted
刚刷到好几个博主都在转,分享个羊毛
上海 AI 实验室的书生·端砚,注册就送每月免费基础包:10 墨点,1 墨点最多抵 5000 万 token,算下来最多 5 个亿
API 重度用户先注册了再说,免费的不薅白不薅
GLM-5.3 Flash Abliterated is FREE 😳
a new research version of GLM-5.3 Flash is now open for testing
- 1M context window
- up to 131K output tokens
- 33.5 tokens/sec
- free to use
- OpenAI-compatible API
built for coding and AI research
available directly in the cloud
test it here: platform.experientiallabs.ai…
free access is available now 👀
已经优化了手机端的适配,发个移动端的视频
几年前,我为了学英语但又不失去乐趣,选择了看英文小说。这个方法来自 B 站 UP 主 @罗肖尼Shawney 。原本以为自己能像看国内网文一样沉迷进去,但实际查单词的痛苦依然历历在目,而且适合自己难度的英文小说本来就很难找。举个例子,英语里一个单词在不同语境、不同年代意思完全不一样,有些俚语因为时代认知偏差,经常会出现“单词都认识,拼在一起却完全读不懂”的情况。
为了解决这个问题,去年的第一个版本是接入 AI,点击单词实时查询。但测试下来体验非常糟糕,熟悉我的朋友可能看到过当时我发的推文,后来也就没正式发布。
不过这一次不一样了。前几天我突然灵光一闪:既然实时查询太慢,为什么不对文章进行通篇预读,让 AI 自动打标、自动解释,直接做成一本配套词典放上去呢?这样就彻底避开了实时查询慢、体验差的问题。
正好今年买了不少 100 美元的 Grok heavy,送了推特蓝V、Cursor 200 plan,还有 Grok bot,算力 Token 根本用不完。我直接在 Bot 上建了 5 个 Agent 拉了个群,把所有 epub 传上去让他们自己分配任务。5 个 AI 老师多线程自动干活,耗时 3 天,烧了一个 Bot 的周限制,项目的第一版终于完成了!
等我把敏感数据抹掉、处理好版权问题后,就会正式开源。
网站 inputread.site
hello retweeted
NEW FREE AI API DROPS
Found more free AI resources worth testing:
1. VyceAI
• Claims daily free credits
• DeepSeek 4.1 Flash
• GPT-6 Astra, claims 25M tokens/day
• Signup required
• Limits not independently verified
2. Apodex 1.1 Mini
• New :free model on OpenRouter
• Rate limited
• Also listed next to Space Bunny Alpha
3. OVHcloud AI Endpoints
• Official free catalog
• Qwen3-Coder-30B
• Qwen2.5-VL-72B
• Mistral Small 3.2
• Anonymous tier is very tight
4. ZyloAI
• Free pool, about 10 RPM
• GLM 5.3
• Kimi K3
• GPT-OSS 20B
• Laguna XS 2.1
• Muse Glimmer 30B
5. Zydit
• Free-models-only endpoint
• DeepSeek R1, Gemma, MiMo, Muse Spark
• 10 RPM
6. GLM-5.3 Flash Abliterated
• Research variant
• 1M context, 131K output
• OpenAI-compatible
• Free for eligible accounts on Experiential Labs
7. APMIX update
• Claims 3M tokens on signup
• Claude Sonnet 4.6 Free on the free tier
• No card
8. GreenPower API
• DeepSeek / GLM / Kimi
• Only $0.20 on signup
• Pay per request after that
Links in the comments 👇
Free APIs keep appearing. 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.
Jackpots on @PlayOnMint are actually more flexible than I thought.
On eligible games, the jackpot isn’t always baked into the normal bet.
There can be a separate jackpot contribution, and when the game supports it, you can choose the contribution level before placing the bet.
That matters because it gives you control over whether you want to participate instead of silently adding it to every spin or wager.
Not every game qualifies either. You have to look for the Jackpot widget or eligibility info inside the game.
And if different contribution levels change your entries or odds, MINT shows that in the live interface instead of using one fixed rule across everything.
Small feature, but I like this kind of transparency.
You can see what you’re opting into, what it costs, and whether the game is actually eligible before you commit.
hello retweeted
神级聚合转换工具 可以把最近撸的所有国模鸡蛋全部聚合
zode的免费GLM
comandcode的1刀套餐
workbuddy国内国际版免费国模
github.com/yetone/magpie 直接把这个仓库克隆在服务器上直接调用
hello retweeted
NEW FREE AI API DROPS
Found more free AI resources worth testing:
1. Kilo Code
• Coding agent with a free model list
• Mostly for experiments, not heavy production
2. OpenCode Zen update
• Fledge Alpha Free
• MiMo-V2.6 Flash Free
• Ling 3.0 / 3.1 Flash Free
• Nemotron 3 Ultra Free
• LongCat 2.5 Preview Free
• Space Bunny still leading usage
3. SambaNova
• Official fast free tier
• Llama and other open models
• OpenAI-compatible
4. Nebius AI Studio
• Free registration tier
• Open models, OpenAI-compatible
5. Nscale
• Free inference tier
• OpenAI-compatible base URL
6. AI21
• Free registration tier
• Jamba / command-style models
7. Z ai official
• New signup pack: claims 20M tokens
• Free Flash models are the 4.x line, not GLM 5.x flagship
• Pack can be used on paid models
8. SiliconFlow
• Chinese gateway
• Free open-model tier
• Qwen / DeepSeek-class models
9. ElectronHub
• Free pool includes DeepSeek V4 Flash
• Codestral, Gemini flash-lite
• Small weekly credit
10. FreeInference
• Claims $20 daily credits
• DeepSeek V4 Flash, GLM 5.3 Flash, MiniMax M3
• 2 concurrent requests
Links in the comments 👇
Free APIs keep appearing. 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.
NEW FREE AI API DROPS
Found more free AI resources worth testing:
1. KiosAPI
• DeepSeek V4 Flash
• DeepSeek V4.1 Flash
• GLM 5.3 Flash
• Claims 30+ more free models
• "No token limits" is a claim, not verified
2. Tiarina Cloud
• Limited-time free flash models
• DeepSeek
• MiMo
• Qwen
• Posted by the founder
3. ZansLab
• Indonesian AI gateway
• Claims free until next week, no top up
• GPT-6 Astra / Sol / Luna
• GPT-5.6 Sol
• DeepSeek 4.1
• Sign up, create a key
4. APMIX event
• Shared pool on GPT-6 Luna
• Model: gpt-6-luna-free
• Works in Claude Code, Cursor, Codex, OpenCode
• Event page, not the old catalog
5. Experiential Labs promo
• GPT-6 Luna
• Qwen 3.8 27B
• DeepSeek V4 Flash
• Jev
• Playground after sign-in
• Older GPT-6 Astra / Fable free promo already ended
Links in the comments 👇
Free APIs keep appearing. 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.
hello retweeted
卧槽,找到这个项目,可以把gpt chat额度给到codex用
试试吧
github万岁
github.com/miuuyy/codex-chat…
hello retweeted
免费不限环境的指纹浏览器 Prism 又更新啦 🚀
🌍 开源社区版现已支持 13 种语言
⚡ Pro 新增 API/MCP 批量创建环境
🎛️ MCP 全局开关,新环境自动继承授权
🖱️ 支持长按、拖拽与按键持续按住
🔗 prismbrowser.app/
#PrismBrowser #MCP #开源 #浏览器自动化
hello retweeted
以后查火车票,可能一句话就够了。
有开发者给 12306 接了 MCP,做了个 12306-mcp,目前已经能直接让 AI 查票。
① 实时查询车票,支持直达和过站
② 按车次、时间、站点筛选
③ 支持查询中转方案
更有意思的是这套 MCP + AI 的玩法。
查票只是开始,未来如果再接上路线、酒店等服务,AI 就能把出行流程串起来。
一句“帮我安排去 XX 的行程”,剩下的交给 Agent。
🔗github.com/Joooook/12306-mcp
hello retweeted
github.com/yinsuecci/nekoai
最近会逐步开源一下自己拿来赚钱的东西
授人以鱼不如授人以渔
hello retweeted
FREE $50 Credits on Render???😳
if you use Render for hosting or not, MLH is giving out free credits through Hacktoberfest 2026
i redeemed mine and got $50.
no need to win a hackathon or do anything, just claim.
that’s free hosting for your side projects, agents, hackathon builds etc.
i’ll drop the redemption link below.
these things usually don’t last forever, so i’d claim it while it’s still available
hello retweeted
体验地址 🫧
idiom.520ai.site/
项目开源地址:
github.com/yuanyang749/softi…
hello retweeted
家里有小朋友的,可以试试这个成语小游戏 🫧
不用逼着背成语,答对一个字,画面、音效和奖励一起冒出来,让孩子在玩的时候顺便记住成语。
抽空做了一个治愈系多巴胺小游戏《成语泡泡》,想把原本有点枯燥的成语学习,变成一件更轻松、更有反馈感的事情。
🎨 高饱和多巴胺视觉
软萌史莱姆 + 纯手绘卡通风格,每次答对都有新的动画反馈。
🎵 爽快正反馈
清脆音效 + 满屏碎纸屑庆祝,让每一次答对都有小小的成就感。
⚡ 双模式设计
🍭 儿童模式:识字启蒙,循序渐进
⚡ 成人模式:高能挑战,测测自己的成语储备
📅 成长与收集玩法
每日签到、今日委托、12 座成语奖杯,把学习过程变成收集和闯关。
纯 Web 前端制作,无需下载,打开就能玩。
喜欢折腾的朋友也欢迎拿去二开,换题库、改玩法,做一个更适合自己孩子的小工具。
如果觉得好玩,求个转评赞 ❤️
如果发现 Bug 或有什么改进建议,评论区直接砸过来~👇
先谢原作者 @grmchn4ai 精妙创意与开源分享
ドパドリル 最开始只是他用 Opus 5.5 做的一个小玩具,没想到开源之后,评论区一直有人催中文版。
所以今晚先把中文界面做出来了,看下面视频效果
这是赶出来的第一版,部分文案可能还没完全替换干净,比如年级名、技能树、提示语之类。如果看到哪里不自然,欢迎直接评论告诉我,我继续改。
15 秒答题,答对一次,画面和音效都会升级一档。越玩越上头,家里有小朋友的可以直接拿去练计算。
体验地址和源码地址在评论区👇
Drex 1.5 is Live! 128k context window, more rivalry to Jev
#1 on the Decision Index 0.2.1 + JevBench (SOTA)
Absolute leader in long context decisions
Latency: sub-second
Decision models are getting better!
Price: $0.04 per 1M input tokens
Try Drex with 250M free tokens
nace.ai/drex
#drex #jev #nace
hello retweeted
上海人工智能实验室有免费的DeepseekFlash
科研导向 比官网API智商还高
每天签到领五亿token
入口:discovery.intern-ai.org.cn/