Same here,, i could not find anywhere on web.

Hi!

Currently, OpenHuman operates on a hybrid approach depending on how you choose to run it:

1. Cloud Subscription: OpenHuman uses an intelligent model routing system that sends tasks to the right commercial LLM (Reasoning, Fast, or Vision) under one unified subscription. Since the project is in early beta, specific subscription tier pricing is managed directly through the desktop application onboarding or within their official community channels.
2. Free / Local Option: If you prefer not to pay a subscription, you can use OpenHuman entirely for free by configuring it to use local AI workloads via Ollama on your own device.

For exact localized pricing or to secure an invite/spot for the cloud subscription, the best place to ask directly is in their community spaces like Discord.

OpenHuman supports custom inference backends through the inference_url and api_key options in config.toml. You can point it at a free API like OpenRouter (no credit card required) and use free models from NVIDIA, Google, DeepSeek etc. instead of the paid subscription.

I wrote a step-by-step setup guide: https://freellm.net/blog/openhuman-model-config-freellm-setup — covers config.toml setup, which free models work best for each task, and how to verify.

The free setup gives you the core experience (desktop agent, memory tree, integrations) without monthly cost.