Team Ai
Apppublic

nghibach/EShip-Coherence

sourceHugging Faceupdated 11d agoView on Hugging Face
0likes
App README

EShip Coherence

The Coherence learning module from EShip Entrepreneurship Studio: canvas blocks, assumption mapping, process mapping, make–buy–partner decisions, revenue–cost analysis, and the nine-element overview.

This is a stateless classroom app. Work remains in the student's browser and can be downloaded with Save file.

Hugging Face configuration

Add OPENROUTER_API_KEY as a Secret in the Space settings.

Optional public variables:

  • —OPENROUTER_MODELS: comma-separated primary and fallback model IDs. Default: openai/gpt-5.6-sol,anthropic/claude-sonnet-5,openai/gpt-5.4.
  • —MAX_OUTPUT_TOKENS: maximum completion budget. Default: 32768; accepted range: 1–131072.
  • —RATE_PER_MINUTE: per-browser AI request limit. Default: 10.
  • —OPENROUTER_MAX_PROMPT_PRICE / OPENROUTER_MAX_COMPLETION_PRICE: maximum USD per million tokens. Defaults: 2.5 / 15.
  • —OPENROUTER_ZDR: set true to require zero-data-retention endpoints. Default: false to preserve the widest failover pool.

OpenRouter provider failover is enabled. GPT-5.6 Sol is primary, Claude Sonnet 5 is the first cross-model fallback, and GPT-5.4 is the final fallback. Routed endpoints must support the requested parameters and must deny data collection. The default price ceiling prevents a temporary promotion ending from silently exceeding the planned price.