Mattbusel/LLMTokenStreamQuantEngine
A low-latency, C++-based simulation engine that ingests token streams from an LLM in real-time, maps semantic token meaning to trade signals, and triggers micro-adjustments to a trading algorithm on a fractional-time (sub-second) scale.
This engine helps quantitative traders turn real-time natural language insights from large language models into immediate, precise trading actions. It ingests a stream of text tokens, translates their semantic meaning into directional trading signals, and then triggers micro-adjustments to a trading algorithm. The ideal user is a quant trader or firm focused on high-frequency, automated trading strategies based on rapid sentiment and market regime shifts.
Use this if you need to integrate LLM-derived market sentiment directly into a high-speed, automated trading system to make sub-second adjustments based on evolving market narratives.
Not ideal if your trading strategies are long-term, not sensitive to micro-second latency, or do not involve real-time natural language processing.
Stars
4
Forks
—
Language
C++
License
—
Category
Last pushed
Mar 18, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/Mattbusel/LLMTokenStreamQuantEngine"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
isEmmanuelOlowe/llm-cost-estimator
Estimating hardware and cloud costs of LLMs and transformer projects
truefoundry/models
Community-maintained registry of AI/LLM model configurations - pricing, features, and limits...
VincenzoManto/llmtrim
A library for trimming tokens in encoding and decoding in LLM (Large Language Model)...
DarkFoot101/Smart-Product-Pricing
Built a multimodal pricing system combining numerical, text, and image features with...