CLI

A small command-line client for chatting with your workspace from the terminal — streaming replies, an interactive session, and stdin piping. It talks to the OpenAI-compatible endpoint, so it needs only a workspace API key.

Install

The CLI ships in the nexevo-ai repository. From a checkout, run it as a module:

bash

python -m cli --help

# optional: a shorter alias for the examples below
alias nexevo='python -m cli'

Configure

Save your workspace API key once (written to ~/.nexevo/config.json, owner-readable only):

bash

nexevo config set api_key sk-ws-...
nexevo config show          # api_key is masked

Environment variables override the saved config: NEXEVO_API_KEY, NEXEVO_BASE_URL (default https://nexevo.ai/v1), NEXEVO_MODEL.

Chat

bash

# one-shot
nexevo chat "Summarize our Q2 results."

# interactive session (keeps the conversation; empty line or Ctrl-D to exit)
nexevo chat

# pipe stdin
git diff | nexevo chat "Review this change."

# pin a specific model (default is nexevo-auto → the router picks)
nexevo chat --model nexevo-auto "Draft a release note."

Replies stream token-by-token. Routing, billing, and BYOK are the same as every other Nexevo call — the CLI is a thin client over the OpenAI-compatible API.