@evsphere/evagent
v0.7.3
Published
ev-agent CLI + evsmem memory server. One-command install, then `evagent -setup`.
Readme
@evsphere/evagent
ev-agent CLI with the evsmem memory server.
Install
npm install -g @evsphere/evagentnpm 12 blocks install scripts by default. If the install warns about blocked
scripts (or evagent reports a missing binary), reinstall with:
npm install -g --allow-scripts=@evsphere/evagent @evsphere/evagentThe evagent command also falls back to the platform package's binary
automatically, so it works even when install scripts were skipped.
Setup (one-time)
Installs evsmem (venv + deps + embedding model) and starts it. Interactive: shows a banner, asks ALL questions upfront (GPU/CPU build, embedding provider, deriver LLM base URL + API key + model name), then installs without further prompts:
evagent -setupUse --auto to skip prompts (downloads BGE-M3 embedding model, ~634 MB):
evagent -setup --autoUse
evagent # interactive TUI
evagent run "msg" # headless one-shotevsmem runs at http://127.0.0.1:9876 (config in evsmem/.env).
