Max K

building @visionagents_ai.

dev advocate @getstream_io, ex: @IBM, @Vonage

public speaker, OS maintainer. python, vision AI, APIs

[Max K @max_does_tech]

legitimately one prompt. you can tell because if i'd used a second one i'd have improved the UI 🤷‍♂️

[Stream @getstream_io]

Stream's agent skills can create an entire messaging app in one prompt!

The skills explain how to build a Stream app to your coding agent.

Prompt it, make some tea, come back to your working app. 🫖

Link in 🧵

[Vision Agents @visionagents_ai]

Real-time avatars are now available in Vision Agents with @Anam__ai to bring custom, responsive experiences to the world!

Here's one cleverly deciding not to trust @stefanjblos with all its company's money... write up in 🧵

[Max K @max_does_tech]

sharing this to get out ahead of the controversy

[Vision Agents @visionagents_ai]

To pineapple or to not pineapple?

@max_does_tech built a voice agent using @inworld_ai's Realtime API to answer the age old question 🍕

Quickstart: https://t.co/DxwUujt9MZ

Example code: https://t.co/VZQPfioVSm

[Max K @max_does_tech]

it could register and track more stuff but i ran out of things to show it

[Vision Agents @visionagents_ai]

This entire example is only 87 lines of code 🤯

This fully-local processor pipeline with @huggingface Transformers object inference and segmentation is running 100% in realtime on a Macbook with the @visionagents_ai SDK!

[Max K @max_does_tech]

i did the huggingface, roboflow and nemotron stuff, maybe more, it's all a blur

[Vision Agents @visionagents_ai]

v0.5.0 of the Vision Agents SDK is out now!

New in this release: run agents directly on your hardware, Anam avatar integration, way faster @DeepgramAI TTS, @AssemblyAI support, and much more.

Details in 🧵👇

[Vision Agents @visionagents_ai]

@Alibaba_Qwen It's amazing how responsive this model is! Really easy to work with too!

https://t.co/x55Ktz0nvI

[Max K @max_does_tech]

for the record, it WAS water

[Vision Agents @visionagents_ai]

Trying the new @Alibaba_Qwen 3.6-Plus model: it's very capable and responsive, not to mention that it takes a lot of creative liberties... 😆

[Vision Agents @visionagents_ai]

Let’s build a vision + voice agent with the new Gemini 3.1 Flash Live model 🔥

Following this tutorial, you'll build a multimodal agent that helps you sell your used items!

Check out the full tutorial on the @googledevs YouTube channel!

https://t.co/AAIJZ4fwaY

[Max K @max_does_tech]

Nice work @stefanjblos! No launch is complete without your dulcet tones

[Vision Agents @visionagents_ai]

Gemini 3.1 Flash Live just dropped, check out our demo with it! 🙌

This @googleaidevs native audio model now comes with lower latency, stronger instruction following and more reliable tool calling.

We'll share the full demo soon!

[Max K @max_does_tech]

Now this one I'm proud of...

[Vision Agents @visionagents_ai]

Using @roboflow's Neural Architecture Search to make a video moderation bot with under 1.8ms average inference time!

In this demo we're able to moderate video coming in on a video call so quickly that you almost don't see the offending content before it's censored 🙌

[Max K @max_does_tech]

wow kimi 2.5 is nailing these evals

[Roboflow @roboflow]

added Cursor cli with Composer 2 to the eval based on community feedback

#2 coding agent for vision tasks

updated blog post if you want to see the details

https://t.co/JORJUC85YT

would like to add more evals, share what issues you're having with vision tasks and we can add them

[Vision Agents @visionagents_ai]

Changelog update time!

We've added:
- @huggingface object detection support
- @AssemblyAI support, including diarization

Upgraded:
- K8s deployment example
- Real time transcript buffering and handling
- SFU error handling

Imminent:
- Non-Stream video options

[Vision Agents @visionagents_ai]

here's @XiaomiMiMo v2-omni being very very polite to me

[Vision Agents @visionagents_ai]

Qwen 3.5 is awesome, it's crazy to have an 0.8B param model that's capable of real time vision understanding!

[Max K @max_does_tech]

It was really satisfying to see my vibe coded frontend actually work

[Vision Agents @visionagents_ai]

Here's @NVIDIAAIDev Nemotron-3-Super-49B used in a real-time Vision Agents application as a fraud assistant! You can see every action it takes, and it's all happening in real-time 😮

Using the Nemotron model hosted on @baseten for reliability.

[Thariq @trq212]

We just added /btw to Claude Code!

Use it to have side chain conversations while Claude is working.

[Max K @max_does_tech]

Editing this one was like living a fever dream

[Vision Agents @visionagents_ai]

v0.4 of Vision Agents is here! Here's the most intense video possible to catch you up on the big hits 🤷‍♂️ Link to the VA GitHub (7K stars, join us!) in bio

[Vision Agents @visionagents_ai]

Interested in computer vision? Here's 31 examples of it in action for ya https://t.co/rAngmZggMf

[Max K @max_does_tech]

This thing is creative in how it wants to screw with me

[Vision Agents @visionagents_ai]

Telling stories with the new GPT-5.4 and got into a fight with it... I want to end the story but it wants to keep playing.

This is a Vision Agents app where the video is being streamed by an SFU to a k8s instance running the actual agent

[Max K @max_does_tech]

actually pretty impressive how responsive this one is

[Vision Agents @visionagents_ai]

This is @GoogleDeepMind Gemini 3.1 Flash-Lite responding in real time in a Vision Agents app. It's able to handle a lot of different video understanding questions much more quickly than the previous gen... and this is on release day, when everyone's hitting the API! 😆