Isra

@Israfilv2 • 18,743 subscribers

AI Founder | Shipped SaveVid AI | i tweets about useful tools, repo and local models | Building cool stuff and worked with 60+ AI & llm brands

Shorts

what the f*ck someone just took the refusals out of Kimi K3.1 Pro 😳 Silk Compute put it on an API and they don't store your prompts or replies -1M context -reads text, images and video -tool calling with max reasoning -works with OpenAI and Anthropic SDKs, just change the base URL -model id: kimi-k3.1-pro-uncensored -top up from $5, no subscription one of the strongest open models with no refusals and no logs. I'd put $5 on it and throw the prompts other models refuse at it try it here 👇

what the f*ck someone just took the refusals out of Kimi K3.1 Pro 😳 Silk Compute put it on an API and they don't store your prompts or replies -1M context -reads text, images and video -tool calling with max reasoning -works with OpenAI and Anthropic SDKs, just change the base URL -model id: kimi-k3.1-pro-uncensored -top up from $5, no subscription one of the strongest open models with no refusals and no logs. I'd put $5 on it and throw the prompts other models refuse at it try it here 👇

48,832 次观看

best AI video so far and it took minutes 😳 Vivix W1 gave me one smooth tracking shot like a real steadicam on a red carpet. glossy leather that catches the light as she walks from the warm chandelier inside to daylight, natural hair movement, jewelry that doesn’t glitch even once and it was fast and surprisingly cheap to make posting this to enter the Vivix Creator Contest with Vivix.AI, go make yours and join too what would you create with it?

best AI video so far and it took minutes 😳 Vivix W1 gave me one smooth tracking shot like a real steadicam on a red carpet. glossy leather that catches the light as she walks from the warm chandelier inside to daylight, natural hair movement, jewelry that doesn’t glitch even once and it was fast and surprisingly cheap to make posting this to enter the Vivix Creator Contest with Vivix.AI, go make yours and join too what would you create with it?

108,186 次观看

what the f*ck, Qwen 3.8 27B just went fully uncensored, and it runs on your gaming pc 😋 -refusals removed, no jailbreak prompts -one 15.7GB file, fits a 16GB GPU -27B dense, MTP for faster replies -30k+ downloads already -open weights. FREEE try it here:

what the f*ck, Qwen 3.8 27B just went fully uncensored, and it runs on your gaming pc 😋 -refusals removed, no jailbreak prompts -one 15.7GB file, fits a 16GB GPU -27B dense, MTP for faster replies -30k+ downloads already -open weights. FREEE try it here:

92,577 次观看

Nace shows Drex playing a strategy game. i gave it a harder opponent: my timeline Spoiler Guard sends each post to the new decision model and blurs the spoilers, 352 post checked in one session with sub-second calls → 128K context window, per Nace the first and only decision model at that size → #1 on the Decision Index, a Jev rival for $0.04 per 1M input tokens → no text generated, just a probability per post try it with 250M free tokens from Nace AI:

Nace shows Drex playing a strategy game. i gave it a harder opponent: my timeline Spoiler Guard sends each post to the new decision model and blurs the spoilers, 352 post checked in one session with sub-second calls → 128K context window, per Nace the first and only decision model at that size → #1 on the Decision Index, a Jev rival for $0.04 per 1M input tokens → no text generated, just a probability per post try it with 250M free tokens from Nace AI:

12,386 次观看

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

nsfw wan 1.3b should be banned, it turns a normal gaming gpu into a dirty video studio 😳 free uncensored text-to-video on hugging face - tuned on adult clips, no built-in filter - 2.8gb file, base model needs about 8gb vram - short 480p clips from just a prompt grab wan_1.3B_exp_e14 and load it in comfyui's wan 2.1 workflow try it here: save this before your gpu finds out what it's really for 18+ only, made-up adults, never real faces

Sensitive content

This media may contain sensitive content.

nsfw wan 1.3b should be banned, it turns a normal gaming gpu into a dirty video studio 😳 free uncensored text-to-video on hugging face - tuned on adult clips, no built-in filter - 2.8gb file, base model needs about 8gb vram - short 480p clips from just a prompt grab wan_1.3B_exp_e14 and load it in comfyui's wan 2.1 workflow try it here: save this before your gpu finds out what it's really for 18+ only, made-up adults, never real faces

236,320 次观看 • 16 天前

this is f*cking insane someone built a browser that uses 16x less memory than chrome 😳 it's open source and made for AI agents -123MB of memory vs chrome's 2GB -100 pages in 5 seconds vs 46 on chrome -works with playwright, puppeteer and MCP -37.8k stars on github if you run AI agents locally, chrome is the thing eating your RAM and this fixes it try it free here 👇

this is f*cking insane someone built a browser that uses 16x less memory than chrome 😳 it's open source and made for AI agents -123MB of memory vs chrome's 2GB -100 pages in 5 seconds vs 46 on chrome -works with playwright, puppeteer and MCP -37.8k stars on github if you run AI agents locally, chrome is the thing eating your RAM and this fixes it try it free here 👇

132,068 次观看 • 12 天前

i don't know why people still are not using this 4b uncensored chat model, only 2.6 GB file, runs on laptop 😳 spark-x2.6-4b abliterated (gguf) - refusals 100/100 down to 2/100 - 1m context still there - tool calls + thinking intact - it has 6177 downloads - apache 2.0, no gate Spark-X2.6-4B-abliterated-Q6_K.gguf (for best quality output use this version) try it here

i don't know why people still are not using this 4b uncensored chat model, only 2.6 GB file, runs on laptop 😳 spark-x2.6-4b abliterated (gguf) - refusals 100/100 down to 2/100 - 1m context still there - tool calls + thinking intact - it has 6177 downloads - apache 2.0, no gate Spark-X2.6-4B-abliterated-Q6_K.gguf (for best quality output use this version) try it here

26,378 次观看 • 3 天前

first time in my life a company is giving me back more than i spent 💀 Drex Max is doing 120% cashback right now. spend $100 on Jev or any other decision model, get $120 back in Drex credits i thought it was a cheap model buying users with promos. it's not: → 78% on OSWorld-Verified, one point behind Opus 5.5 → a web task that costs $0.074 on Claude Code costs $0.015 on Drex → API is Jev compatible, so you only change 3 env vars almost same score, 5x cheaper, and you get paid to try it $25 free credits at + use code "ISRA10" for $10 more try here:

first time in my life a company is giving me back more than i spent 💀 Drex Max is doing 120% cashback right now. spend $100 on Jev or any other decision model, get $120 back in Drex credits i thought it was a cheap model buying users with promos. it's not: → 78% on OSWorld-Verified, one point behind Opus 5.5 → a web task that costs $0.074 on Claude Code costs $0.015 on Drex → API is Jev compatible, so you only change 3 env vars almost same score, 5x cheaper, and you get paid to try it $25 free credits at + use code "ISRA10" for $10 more try here:

11,592 次观看 • 7 天前

i gave MiniMax M3.1 Flash one prompt and it built me a whole operating system in a single HTML file 😳 here's what it pulled off: → frontend: boot screen, lock screen, dock and spotlight search, all clean and smooth → complex logic: a window manager where you can drag, resize, snap and minimize → full apps: terminal, files, notes, paint, music and snake, all sharing one file system → 3D: a galaxy app built in Three.js → agentic tool use: an AI assistant inside the OS, running on M3.1 Flash itself, that opens apps and creates files for you took 20 minutes and cost me $0 extra (M3.1 Flash is unlimited for MiniMax Code subscribers till oct 25) 💀 prompt in the replies

i gave MiniMax M3.1 Flash one prompt and it built me a whole operating system in a single HTML file 😳 here's what it pulled off: → frontend: boot screen, lock screen, dock and spotlight search, all clean and smooth → complex logic: a window manager where you can drag, resize, snap and minimize → full apps: terminal, files, notes, paint, music and snake, all sharing one file system → 3D: a galaxy app built in Three.js → agentic tool use: an AI assistant inside the OS, running on M3.1 Flash itself, that opens apps and creates files for you took 20 minutes and cost me $0 extra (M3.1 Flash is unlimited for MiniMax Code subscribers till oct 25) 💀 prompt in the replies

25,234 次观看 • 20 天前

what the hell did this free stealth model just build 💀 I asked Space Bunny, a Flash-level model, to make Minecraft in one HTML file and left it running it came back with a playable survival game inside my browser I could explore the generated world, break and place blocks, and switch items through the hotbar easily the best result I have got from a coding model on this test Space Bunny is free on OpenCode for the next week prompt below:

what the hell did this free stealth model just build 💀 I asked Space Bunny, a Flash-level model, to make Minecraft in one HTML file and left it running it came back with a playable survival game inside my browser I could explore the generated world, break and place blocks, and switch items through the hotbar easily the best result I have got from a coding model on this test Space Bunny is free on OpenCode for the next week prompt below:

44,311 次观看 • 1 个月前

I STOPPED PAYING FOR CODING MODELS THIS WEEK 💀 Antseed has a desktop app that works like a VPN for AI. turn it on, pick a model, and it routes you to whoever is selling it cheapest deepseek v4 flash, minimax m3, gpt oss 120b and qwen3 235b is all free setup: → download from → activate the vpn → pick a model, or point codex/claude code at it no account, no email needed thank me later

I STOPPED PAYING FOR CODING MODELS THIS WEEK 💀 Antseed has a desktop app that works like a VPN for AI. turn it on, pick a model, and it routes you to whoever is selling it cheapest deepseek v4 flash, minimax m3, gpt oss 120b and qwen3 235b is all free setup: → download from → activate the vpn → pick a model, or point codex/claude code at it no account, no email needed thank me later

31,248 次观看 • 1 个月前

Your coding agent can burn the whole team budget overnight Antseed Gateway fixes that. It's a self hosted OpenAI compatible endpoint with a spending cap and usage tracking on every key I set one up, gave Cline its own $1 a day key and ran a free model through it. Every call was logged and my teammate's key stayed untouched What I like most is that the agent gets its own budget, so I'm not just trusting it to behave

Your coding agent can burn the whole team budget overnight Antseed Gateway fixes that. It's a self hosted OpenAI compatible endpoint with a spending cap and usage tracking on every key I set one up, gave Cline its own $1 a day key and ran a free model through it. Every call was logged and my teammate's key stayed untouched What I like most is that the agent gets its own budget, so I'm not just trusting it to behave

17,022 次观看 • 29 天前

MINIMAX OPEN SOURCED A CODING AGENT THAT LIVES IN YOUR TERMINAL its called MiniMax Code CLI. i gave it one prompt, 19 minutes later i had a working interactive 3d globe. → 249 countries → 14 animated flight arcs across 16 hubs → click a country and the camera flies to it → fresnel atmosphere, bloom, the full pass it did it in 9 phases at about 96 tok/s, committing after every one. Mcode is MIT and it is not locked to MiniMax models, you can bring your own key and point it at claude, gpt, or a local model. setup in your terminal: 1.⁠ ⁠curl -fsSL | bash 2.⁠ ⁠cd into the project and run mcode 3.⁠ ⁠run mcode init . first, that is the step that writes the AGENTS.md the AGENTS.md is doing more work here than the model. build commands, what not to touch, and a rule that it verifies before it calls anything done i will drop the repo below

MINIMAX OPEN SOURCED A CODING AGENT THAT LIVES IN YOUR TERMINAL its called MiniMax Code CLI. i gave it one prompt, 19 minutes later i had a working interactive 3d globe. → 249 countries → 14 animated flight arcs across 16 hubs → click a country and the camera flies to it → fresnel atmosphere, bloom, the full pass it did it in 9 phases at about 96 tok/s, committing after every one. Mcode is MIT and it is not locked to MiniMax models, you can bring your own key and point it at claude, gpt, or a local model. setup in your terminal: 1.⁠ ⁠curl -fsSL | bash 2.⁠ ⁠cd into the project and run mcode 3.⁠ ⁠run mcode init . first, that is the step that writes the AGENTS.md the AGENTS.md is doing more work here than the model. build commands, what not to touch, and a rule that it verifies before it calls anything done i will drop the repo below

25,336 次观看 • 1 个月前

this AI saw a guy in the background and chose violence Tavus gave me early access to griffin and this clip had me dying. dude walks behind the laptop in a thrifted sweater and it roasts him on the spot i've used a lot of AI video stuff and none of it ever noticed anything i didn't say first. this one just looked up and reacted #1 on NVIDIA's full-duplex benchmark, and this is only the lite model @hassaanrza's team shipped

this AI saw a guy in the background and chose violence Tavus gave me early access to griffin and this clip had me dying. dude walks behind the laptop in a thrifted sweater and it roasts him on the spot i've used a lot of AI video stuff and none of it ever noticed anything i didn't say first. this one just looked up and reacted #1 on NVIDIA's full-duplex benchmark, and this is only the lite model @hassaanrza's team shipped

20,412 次观看 • 1 个月前

i'll be honest, i expected "AI-made game" to mean janky demo then i played Farmadillo by Loccle 🖍️ from the Sandbox beta jam winners and the farm loop actually got me, seeding crops, adding a bee house, then a chicken house, weirdly satisfying it's a farming sim crossed with pinball and tower defense btw, your armadillo companion defends your crops Sandbox studio in beta already producing this is a good sign play it:

i'll be honest, i expected "AI-made game" to mean janky demo then i played Farmadillo by Loccle 🖍️ from the Sandbox beta jam winners and the farm loop actually got me, seeding crops, adding a bee house, then a chicken house, weirdly satisfying it's a farming sim crossed with pinball and tower defense btw, your armadillo companion defends your crops Sandbox studio in beta already producing this is a good sign play it:

16,258 次观看 • 1 个月前

1M free tokens every month for Opus 5.5, GPT-5.6, Grok 4.6, DeepSeek V4 and 29 more models with zero card 😳 I tested every model this account can access. all 33 answered through one OpenAI-compatible API what you get for free: - 1M tokens every month - 60 requests per minute - Claude, GPT, Grok, Gemini, DeepSeek, Qwen, Kimi and GLM models - works with Cursor, Cline, opencode, Claude Code, Continue and Aider setup: 1. go to 2. create your free account 3. generate your own API key 4. set the base URL to 5. pick a model and test one call CodeCraft is a third-party gateway, so use it for testing and keep private work off it bookmark this before the model list changes

1M free tokens every month for Opus 5.5, GPT-5.6, Grok 4.6, DeepSeek V4 and 29 more models with zero card 😳 I tested every model this account can access. all 33 answered through one OpenAI-compatible API what you get for free: - 1M tokens every month - 60 requests per minute - Claude, GPT, Grok, Gemini, DeepSeek, Qwen, Kimi and GLM models - works with Cursor, Cline, opencode, Claude Code, Continue and Aider setup: 1. go to 2. create your free account 3. generate your own API key 4. set the base URL to 5. pick a model and test one call CodeCraft is a third-party gateway, so use it for testing and keep private work off it bookmark this before the model list changes

14,705 次观看 • 1 个月前

没有更多内容可加载