Files
2026-07-13 12:58:18 +08:00

40 lines
1.7 KiB
Bash

# API Keys (shared across integrations)
OPENAI_API_KEY=replace-with-your-key
ANTHROPIC_API_KEY=replace-with-your-key
# Agent backend URL (for the CopilotKit runtime proxy)
AGENT_URL=http://localhost:8000
# Showcase
NEXT_PUBLIC_BASE_URL=http://localhost:3000
# --- aimock (dev / CI / smoke) ---
# Point the Python agent at a local aimock server instead of real OpenAI.
# Leave unset in production. When set, OPENAI_BASE_URL + LITELLM_API_BASE are
# populated automatically and a dummy OPENAI_API_KEY is injected if absent.
#
# The `/v1` suffix is REQUIRED — the OpenAI SDK appends its own paths beneath
# whatever you provide here. Port 4010 matches what CI uses; pick any free
# port locally.
#
# Start aimock locally with: npx aimock --port 4010
# AIMOCK_URL=http://localhost:4010/v1
#
# USE_AIMOCK controls the toggle when AIMOCK_URL is also set:
# (unset/empty) — enabled (default when AIMOCK_URL is set)
# 1, true, yes, on — enabled
# 0, false, no, off — explicitly disabled (keeps AIMOCK_URL intact)
# anything else — disabled with a warning
# Setting USE_AIMOCK=1 with no AIMOCK_URL is a no-op (warning logged).
# USE_AIMOCK=1
#
# Production guard: if ANY of the following env vars is set to 'production'
# or 'prod' (case/whitespace tolerant), aimock_toggle REFUSES to apply even
# when AIMOCK_URL is set, and logs a WARNING explaining why:
# NODE_ENV, ENV, ENVIRONMENT, APP_ENV, DEPLOY_ENV, PYTHON_ENV,
# RAILWAY_ENVIRONMENT, RAILWAY_ENVIRONMENT_NAME
# This prevents a misconfigured prod deploy from silently redirecting real
# traffic through a mock server. To keep the guard from firing in dev/CI,
# leave these unset or set them to something other than 'production'/'prod'
# (e.g. 'development', 'preview', 'staging').