hosted.env.example

# Hosted fallback on Azure AI Foundry - use ONLY if your local model will not run.
#
#   1. Copy this file to env/hosted.env
#   2. Replace the two <...> values with the endpoint and key from your workshop instructors
#   3. Carry on with the workshop exactly as everyone else does
#
# This file overrides the model settings in every step, so you never edit it again.
# env/hosted.env is git-ignored - never commit a key.
#
# The deployment serves Qwen3-4B-Instruct-2507 - the same model as the recommended local tier
# (qwen3:4b-instruct) - so no exercise puts you at a disadvantage. Qwen3-1.7B isn't offered as a
# hosted model in Azure AI Foundry, which is why the fallback uses the 4B instruct model.
MODEL=openai-chat:qwen3-4b-instruct
OPENAI_BASE_URL=https://<your-foundry-resource>.services.ai.azure.com/openai/v1/
OPENAI_API_KEY=<your-api-key>
REASONING=default
#
# The endpoint also speaks the Responses API: MODEL=openai-responses:qwen3-4b-instruct works too.
# openai-chat: is the default because it measured faster on this deployment (about 1.8 s per turn
# against 3.2 s).
#
# Running the workshop yourself? scripts/hosted-fallback.ps1 (or .sh) deploys the model to your own
# Azure AI Foundry resource ("up", 15-20 minutes), writes this file for you ("env"), and removes the
# deployment afterwards ("down"). Managed compute bills per GPU-hour while the deployment exists
# (about USD 3.95/hour for one A100), so run "down" as soon as the session ends.
#
# Any other OpenAI-compatible endpoint works too: change the model, URL and key above to its values.