Consistency Python Packages

rerun-bench

Same task, run N times: how consistent and how expensive is your coding agent? A rerun benchmark for Claude Code, Codex and OpenCode.

rankseg

Boost segmentation model mIoU/Dice instantly WITHOUT retraining. A plug-and-play, training-free optimization module. Published in NeurIPS & JMLR. Compatible with SAM, DeepLab, SegFormer, and more. 🧩

timbro

Keep your writing sounding like you, even when an LLM is doing the writing. Local, white-box voice-distance scoring that returns named, content-preserving edits. MCP-ready.

agent-convergence-scorer

CLI and Python library for lexical stability checks on repeated short agent outputs: pairwise exact agreement, token overlap, divergence point, and an order-independent score. Zero runtime dependencies.

engram-team

Agent memory platform.

snipwise

Keep text snippets in sync across files in a project

llm-testlab

Comprehensive Testing Tool for Large Language Models

memstate

Transactional Memory for AI Agents - Keep SQL and Vector DBs in sync with ACID-like guarantees

worldcut

Cross-service consistency verification for autonomous agent decisions

agentcrdt

CRDTs for distributed AI agent state and conflict-free belief merging

mlscorecheck

Testing the consistency of binary classification performance scores reported in papers

texttoknowledge

Convert unstructured text into structured, queryable knowledge with llmatch-messages. Extract and organize key details for fast, reliable access—ideal for teams and researchers.

partnershipparser

partnershipparser extracts and structures key info from tech partnership news for easy analysis of companies, focus areas, and impacts

headline-parser

A new package that processes news headlines or short text snippets to generate structured summaries of current events. It uses an LLM to extract key entities, topics, and sentiment, ensuring the outpu

data-convertible

A new package that provides a structured and reliable way to process user input related to common developer utilities such as JSON, Base64, URL, and hash operations. It uses an LLM to interpret user r

structuredxtract

A new package is designed to simplify the process of extracting structured information from user-provided text inputs by leveraging a language model with pattern matching capabilities. The system prom

scamp-analyzer

A new package that analyzes user-provided text descriptions of mischievous or cleverly unconventional behavior (a "scamp" scenario) and returns a structured assessment. It evaluates the creativity, hu

sport-evo-summary

A new package that transforms unstructured text about sports evolution into a structured summary. Users input text describing changes in a sport, and the package returns a standardized breakdown of ke

pre-v7-unix-summarizer

A new package that processes text input related to early Unix history (pre-V7) and returns structured, verified summaries using pattern matching and LLM interactions. It takes user-provided text (e.g.

leaderboard-summarizer

Extracts and organizes leaderboard/ranking data from text via smart pattern matching and retries, delivering clean, structured summaries effortlessly.

text-to-struct

text-to-struct converts unstructured text into standardized, structured output

vrscene-parser

A new package would process textual descriptions of virtual reality scenes or environments and return structured, validated outputs that describe the scene in a standardized format. It would use an LL

thematic-structurizer

A new package would process user-provided text input related to historical or thematic content—such as summaries, descriptions, or analyses of topics like persuasion techniques from antiquity—and retu

vitae-parser

vitae-parser extracts and structures biographical details from unstructured text using pattern matching.