wehub-resource-sync
16 lines
810 B
TOML
16 lines
810 B
TOML
name = "Requesty"
|
|
env = ["REQUESTY_API_KEY"]
|
|
npm = "@ai-sdk/openai-compatible"
|
|
api = "https://router.requesty.ai/v1"
|
|
# Reasoning HTTP format (accessed 2026-06-25):
|
|
# POST /v1/chat/completions uses top-level reasoning_effort as an effort string
|
|
# or decimal budget string. OpenAI: low|medium|high pass through, max -> high,
|
|
# none|min -> low; budgets 0..1024/1025..8192/>=8193 map low/medium/high.
|
|
# Anthropic: none|min|low -> 1024, medium -> 8192, high -> 16384, max ->
|
|
# model max output minus 1. Vertex: none|min -> 0 for Flash/Flash Lite or 128
|
|
# for Pro, low -> 1024, medium -> 8192, high -> 24576, max -> max output.
|
|
# Thus none is a gateway translation and does not always disable reasoning.
|
|
# Source:
|
|
# https://docs.requesty.ai/features/reasoning
|
|
doc = "https://requesty.ai/solution/llm-routing/models"
|