Ariska (@ariskaa_ai) — sponsored posts & rates

unfair.so

Browse more creators

@ariskaa_ai

Ariska11.8K

Ariska

This creator added themselves. Not verified by unfair.so.

@ariskaa_aiAI ToolsAI CodingAI Agents

26 | writing about AI, automation, tools & growth for builders & brands

11.8KFollowers

15.5KMedian impr. / post

0.3%Engagement

Paid partnerships

In their audience

11+ affiliated with @weteamng follow Ariska

@Rabiussunny11

@SazidOnX

@Mehedi_ng

@Gudia871

@jontu51

@JisanX_Dolphin+5

5+ affiliated with @axisrobotics

4+ affiliated with @fcbarcelona

3+ affiliated with @okx

3+ affiliated with @manutd

3+ affiliated with @chelseafc

1,500+ verified followers analyzed

@Rabiussunny11

@SazidOnX

@Mehedi_ng

@Gudia871

@jontu51

@JisanX_Dolphin+1494220+ media creators85+ crypto creators44+ AI creators33+ creators30+ engineers29+ crypto marketers

25–34 core age group

41.3%

13.6% of audience in India

🇮🇳

Last 7 days62 posts analyzed · replies not counted · updates weekly

ImpressionsEngagements

329K164K0

2.4K1.2K0

Aug 30Aug 31Sep 1Sep 2Sep 3Sep 4Sep 5

Past collabs

Ariska@ariskaa_ai · Sep 6OpenAi's Astra is running the same recurrent direction sapient already shipped with HRM, only you can't look inside astra HRM-Text has been fully open since may, and i mean fully: → the architecture → the weights → the entire pretraining framework → ~1B params, 40B tokens, 8xH100, 50 hours, all documented so with astra you just trust the description, with HRM-Text i can clone the repo tonight, see exactly how it trains and build on the same direction myself → http://github.com/sapientinc/HRM-Text

SapientPaid partnership3.3K views

Ariska@ariskaa_ai · Sep 2i built a full playable browser game in MiniMax Code. no template, no cloning, no copy-pasted assets. one chat thread start to finish + H3 video model for the cinematics what i built, sketchbook bosses (3-stage boss fight): → summon: pick 1 of 3 hand-drawn doodle bosses on the page → fight: top-down arena, WASD + space + click, dodge ink bombs, telegraph rings, melee charges → seal: eraser smash in the arena, then H3 ending cinematic 3 short H3 videos baked in (opening / reveal / ending) with matching audio, 2K, 15s each how i built it: → described the game in one prompt → Mcode wrote the JS, CSS, and doodle boss SVG art → asked for 3 H3 clips and it kept the same look across all of them, same sketchbook, same spiky orb, same paper texture, didn't expect that → tested inside Mcode's built-in browser → found a layering bug in the result screen, told it to fix, fixed on next pass → shipped as a static page cost + workflow: → 3 H3 videos ran about 4600 credits (15s × 3) → whole build stayed in one chat thread, no jumping between separate tools for video, code, images, deploy → the test + edit loop made it feel like actual game dev, not a one-shot demo what H3 inside Mcode actually does: → H3 is MiniMax's new video model, built right into Mcode → same character showed up looking the same across 3 separate generations without me describing it again → the loop: watch the video, tell Mcode what to change, get a new one daily check-in is live, 400 pts/day, 1000 on day 4 & 7, 4000 a week, i used mine on those H3 clips above → http://agent.minimax.io/download

MinimaxPaid partnership88.9K views

Ariska@ariskaa_ai · Sep 1karpathy wrote 630 lines of python, went to sleep, and woke up to 50 finished machine learning experiments. one gpu, no human. but the only thing checking those results was the same AI that produced them. its like a chef doing his own food inspection. so @EvoMapAI just open sourced AutoResearch to fix exactly this ↓

EvomapPaid partnership40.3K views

Ariska@ariskaa_ai · Sep 1100M Freeee tokens for grok-4.5, gpt 5.6 sol and grok-4.6 today (also many other models) 😳 Api key: pkay_3a6dd5ce9757fa34fb8949d511148dbf77bba7860d2dedcd base url: https://api.pkay.fun/v1 endpoint: https://api.pkay.fun/v1/chat/completions openai-compatible, so it drops straight into claude code, codex, cursor how to use it: > export base url + key into your cli > pick model "grok-4.6" or "grok-4.5" and run models i also saw in their /models endpoint (i tested it just now): claude opus 5, claude opus 4-8, claude sonnet 5, claude haiku 4.5 gpt-5.6-sol deepseek v4 flash, deepseek v4 pro kimi k3, kimi k2.6 minimax-m3 llama 4 maverick, gemma 4 26b glm-5.2 gemini 3.6 flash third-party gateway, not the official xAI endpoint. free tier limits can shift without notice, so run it on junk projects first and keep your important keys on the official keys. save this if you wanted grok-4.6 without paying xai $200/mo.

PkayPaid partnership20.8K views

Ariska@ariskaa_ai · Aug 31someone built a live directory of ai coding free api credit sites http://github.com/panxunying/ai-coding-welfare it tracks free/proxy sites you can plug into claude code, codex, and cursor what it does: - lists sites with signup credits and daily check-ins - auto-updates credits, models, prices, and online status - one-click setup scripts for mac, linux, and windows - base urls + env vars ready for claude code sites currently tracked (as of aug 30): agentrouter, justdowork, rawchat, matrix, tabiai, gorouter example first-day numbers they list: agentrouter ~$175 tabiai ~$120 justdowork ~$92 gorouter ~$70 models they show on agentrouter: claude opus 5, opus 4.8, glm-5.3, deepseek-v4-flash, gpt-5.6-sol setup: 1. open the repo 2. pick a site from the table 3. register and make an api key 4. run their quickstart script, or paste the base url into claude code these are third party relays, not official anthropic or openai don't send private code or secrets through them save the repo if you want one place that updates these free coding endpoints for you

GithubPaid partnership42.4K views

Planning a launch?Ariska is one of 200 curated B2B creators. Tell us what you are launching and we will put the shortlist together.

[email protected]·t.me/fdrtrm

1 picked$180