inference.sh
@inference_sh
Followers
145
Following
44
Media
5
Statuses
16
a unified platform to explore, build, run, and scale ai workflows on consumer hardware, on-premise servers, or in the cloud.
San Francisco, CA
Joined July 2024
vibe-building a full game of werewolf. my cofounder thinks this is the best way to show deep-agents, parallel tool calls, and on-the-fly ui generation
1
4
10
and of course it's compatible with all the themes by default
deep agents, parallel tools, ui on the fly built in a few clicks. runs anywhere. api-ready hold on to your prompts, more is coming @inference_sh
0
1
4
👀👀
deep agents, parallel tools, ui on the fly built in a few clicks. runs anywhere. api-ready hold on to your prompts, more is coming @inference_sh
0
0
2
FLUX.2 just landed on @inference_sh. drop it into your workflows, wire it into agents, chain it with anything. go wild.
FLUX.2 is here - our most capable image generation & editing model to date. Multi-reference. 4MP. Production-ready. Open weights. Into the new.
1
1
9
text to research + image + audio + video. meet the bay: a new https://t.co/UMC0v3om9G layer that does things. media, text, api calls, vibed one-off code. it's wild. no limits. local or cloud. self-hosted version baking. lmk if you want to see this before others.
5
2
6
A few things I need to get off my chest about the V4 launch. This morning, when we wanted to release, one of our GPU clusters had a meltdown (still can't believe that timing...) Thanks to unreal INSTANT (5s) reaction and support from our GPU partner @gcore_official , the
3
3
22
â–³ new model: mmaudio add audio to any video run local or on cloud now: https://t.co/bGVOxxZK3K
10
1
16
â–³ new model: higgs-audio generate conversations based on text input run local or on cloud now: https://t.co/dR9BGiDep6
3
1
5
Absolutely and beautifully distracted by this
was optimizing Flux for @inference_sh but fr got sidetracked making svg animations for that last 3.9 seconds.
2
1
17