AgentsOpen sourceMIT
FinMem
LLM trading agent with layered memory and a tunable risk character, from the paper of the same name
By Yangyang Yu et al.
Live GitHub metrics
Live GitHub data · Oct 11, 2026Stars
963
Stars · 7d
+3
Stars · 30d
+10
30d momentum
+1.0%
Forks
195
Open issues
8
Language
Python
Last push
2 years ago
No tagged releases. Repository created Dec 1, 2023.
What it is
FinMem adds a profiling module, a layered memory (short, mid and long term) and a decision module to an LLM so that it can trade single stocks from news and prices. The repository is the research code for the paper; it is not maintained as a product.
Research code accompanying a paper; last pushed in 2024.
In the maintainers' words: “FinMem: A Performance-Enhanced LLM Trading Agent with Layered Memory and Character Design”
Markets
- US equities
Tags
- Research code
- Memory
- Single-stock