Ai Research Python Packages

scitex

Python toolkit for reproducible science — from raw data to manuscript. Includes 42 modules, 312 CLI commands, 0 MCP tools, and 0 skills.

scitex-dataset

Multi-domain scientific dataset fetcher — neuroscience, biology, pharmacology, medical. Part of SciTeX.

halftrace

Diagnostics for agent loops. Measure how fast each capability decays over long trajectories.

inflion

Observability and influence tracing infrastructure for multi-agent AI systems.

autocrucible

Autonomous experiment platform — let LLM agents optimize anything through iterative edit-evaluate loops

rlm-code

RLM Code: Research Playground & Evaluation OS for Recursive Language Model Agentic Systems

notebooklm-skill

NotebookLM does the research, Claude writes the content. Research → Synthesis → Content Creation → Publishing. Claude Code Skill + MCP Server.

swarms-cloud

Deploy your autonomous agents to production grade environments with 99% Uptime Guarantee, Infinite Scalability, and self-healing.

llm-tokens-atlas

Open benchmark of LLM tokenization calibration across providers.

nouse

A persistent epistemic substrate for AI. Nous treats language models as larynx, not mind, and benchmarks epistemic structure instead of output fluency.

hn-tech-signal-mcp

MCP server for global tech & AI signal intelligence via HackerNews, arXiv, Lobste.rs and GitHub

cogitator

A Python toolkit for chain-of-thought prompting 🐍

autoevolve

Multi-agent research competition orchestrator for autoresearch

traceiq

Measure AI-to-AI influence in multi-agent systems

autojudge

Smarter experiment evaluation for autoresearch — replaces eyeballing val_bpb with statistical verdicts

autosteer

Companion tools for Karpathy's autoresearch - smarter evaluation, guided steering, and multi-agent competitions for GPT pretraining

zotero-keeper

MCP Server for managing local Zotero libraries - enables AI assistants to read and write bibliographic references

cortex-research-suite

[Legacy] AI Research OS — 27 self-evolving skills. Decomposed into 20 repos at github.com/TECHKNOWMAD-LABS

iris-gate

IRIS Gate Protocol: Cross-architecture phenomenological convergence research (RFC v0.2)

vortex-fusion

Transformers + Mambas + LSTMS All in One Model

theorist

Prediction-error optimization. The agent guesses before each experiment. Wrong predictions are the learning signal. One file, zero dependencies. pip install theorist