a small studio · est. 2026

wehuman

three tools, one stance

All three do the same thing: the machine carries the repetition, the person keeps the controls.

awedot

  • sessions
  • project context
  • multi-agent

GitHub ↗

awedot.wehuman.top/ ↗

When attention moves on, an agent session should not vanish with it. awedot saves, organises, restores — session memory, project context, multiple agents.

boundaries
  • Sessions stay on your machine; awedot is a local index, not a cloud.
  • Restores are explicit — nothing auto-resumes.

awewarm

  • plans ready
  • local-first
  • explicit boundaries

GitHub ↗

While you are away, coding-plan windows go cold. awewarm fires one minimal scheduled request in your working hours, so they are warm when you return.

boundaries
  • Your schedule lives in a local config, nowhere else.
  • The boundary is explicit: one small request at a planned time, nothing more.

aweshare

  • local models
  • self-hosted hub
  • standard SDKs

GitHub ↗

aweshare.wehuman.top/ ↗

A local-first relay for AI capability. Friends point ordinary OpenAI / Anthropic SDKs at a hub; keys are injected at call time, on the provider’s machine, and never travel.

boundaries
  • Stated plainly: prompts and responses pass the trusted hub in cleartext — not end-to-end encryption.
  • Everyone still honours the upstream services’ terms.
  • The hub can be yours (aweshare hub serve); either way, keys stay home.