
by FutureHouse
Aviary is an open-source gymnasium from FutureHouse for defining custom language-agent reinforcement learning environments. It ships with environments for math, general knowledge, biological sequences and scientific literature search.
AgentFree
Step 1: Install the library with pip install fhaviary.
Step 2: Add extras such as gsm8k or labbench if you want the existing environments.
Step 3: Subclass Environment and define reset and step with your tools.
Step 4: Pair it with an LDP agent and sample trajectories with a rollout manager.
Step 5: Follow the tutorial notebooks to build and evaluate your own environment.
It is a class with reset and step methods that gives an agent messages and tools, then returns observations and a reward after each action.
The readme lists math, general knowledge, biological sequences, scientific literature search and protein stability, plus a notebook environment.
It is Aviary's sister library for defining language agents as Language Decision Processes and running them on Aviary environments.
Run pip install fhaviary, or add extras in brackets to include specific environments.
FutureHouse PlatformScientific-agent platform for literature search, deep research, evidence synthesis, novelty checking, and chemistry and biology research workflows
PaperQA2PaperQA2 is a Python package for retrieval augmented question answering over scientific papers and other documents. It gives answers with in-text citations and can run as an agent that refines its own searches.
Devv AIAI developer search and coding assistant for technical research, code questions, repository context, and implementation guidance
BioMCPBioMCP is a single-binary command-line tool and MCP server for querying biomedical databases. It gives researchers, clinicians and agents one command grammar instead of many sites and APIs.