wehub-resource-sync
17 lines
938 B
TOML
17 lines
938 B
TOML
name = "Neuralwatt"
|
|
env = ["NEURALWATT_API_KEY"]
|
|
npm = "@ai-sdk/openai-compatible"
|
|
# Raw HTTP is POST `/v1/chat/completions`. Reasoning models accept top-level
|
|
# `thinking_token_budget` (no bounds or disable sentinel documented). Native
|
|
# toggles use `chat_template_kwargs.enable_thinking = true|false`; GLM-5.2 also
|
|
# accepts `reasoning_effort`, with the model-specific normalization documented.
|
|
# Flex tier uses either a `-flex` model ID or `service_tier = "flex"` and
|
|
# requires `stream = true`; non-streaming requests fall through to the standard
|
|
# tier. The official Energy Methodology docs currently state a 0.5 flex pricing
|
|
# multiplier.
|
|
# https://portal.neuralwatt.com/docs/api/chat-completions (accessed 2026-06-25)
|
|
# https://portal.neuralwatt.com/docs/guides/flex-tier
|
|
# https://portal.neuralwatt.com/docs/energy-methodology (accessed 2026-06-28)
|
|
api = "https://api.neuralwatt.com/v1"
|
|
doc = "https://portal.neuralwatt.com/docs"
|