The benchmark is in

2.3× faster
36% cheaper
7.7× less memory
8× less CPU

more info: github.com/tinyhumansai/o…

bro casually became the #1 most active @github contributor in India

that's our founder @senamakel btw.

some people talk about building. bro IS building.

absolute machine. 🫡

OpenHuman setup in UNDER 60 SECONDS.

compare this with other agent harnesses?

get started: tinyhumans.ai/openhuman

Fastest growing. Cheapest to run. Best in the game - OpenHuman's growth @github 🐥

Stars going up.
Forks going crazy.
Costs going down.

#opensource #aiagent

OpenHuman is the fastest growing opensource agent harness

github.com/tinyhumansai/o… (github updated)

AGI before flying cars is crazy

if you can open-source code, why can't you open-source intelligence?

trending on @X for the right reasons

Besides that, we also gave OpenHuman a serious quality-of-life upgrade.

• files and agent outputs are easier to find and manage
• smarter chat and agent workflows
• better tool calling and integrations
• more reliable search and memory module loading

less debugging. less ing your AI do its thing.

OpenHuman v0.64.15 is now live, with the latest fixes on top of our memory and voice upgrade.

open source. yours to run.

tinyhumans.ai/openhuman

OpenHuman new release! v 0.64.15

Your AI now remembers more, talks back and actually gets better at working with you.

New memory engine. Live voice agents. free memory inference. better workflows.

And a whole lot of fixes under the hood.

Here's everything we just shipped 🧵

Voice Agents are now live!

You can now have real-time conversations with your agent using supported voice services.

Ask questions. get spoken responses. interact naturally.

Currently in alpha and we're just getting started. 🎙️

And yes memory inference is on us! You can ingest whatever you want and memory storage is based on your plan.

Basic users get 1 GB of memory storage. Pro users get 20 GB.

You can even self-host CortexDB yourself and keep everything private.

We like that math.

OpenHuman's memory now runs on @cortexdbai (and we're sunsetting tinycortex 🪦)

There are easy to use migration helpers to help you migrate.

You now get better/faster recall, smarter ingestion and free storage.

More memory engines incoming...

80,396 humans later… tiny isn’t so tiny anymore 💛

OpenHuman did 802M tokens on @OpenRouter yesterday

+74% in a day

OpenHuman casually back on github trending 🫡🫡

what’s one thing you’ve tried getting an AI agent to do that it still completely sucks at?

we'll build it for you inside OpenHuman