Self-healing PDF extraction that flags what it can't read — and now certifies any extractor's output, catching silently-dropped pages. Free, MIT. 7-tool MCP server. Patent-pending method.

Extract structured data from local or remote LLM models

Intelligent document OCR and structured extraction. Turn PDFs and images into typed JSON with zero vendor lock-in.

Local-first eval harness for unstructured-document extraction — compare LLMs, OCR/IDP tools, and strategies (cascade/ensemble/verify) on the same cohort.

A schema-driven framework for LLM structured extraction enhanced by multi-stage RL training (SFT→DPO→GRPO), with interpretable reward design and end-to-end reproducibility.

Memtruth SDK: evidence, parse, corpus contracts, chunking, projection, and diagnostics for AI applications.

Claude Code Skill for structured information extraction from code/docs/logs. 6-step Python pipeline (source grounding, dedup, confidence scoring, entity resolution, relation inference, KG injection). Zero dependencies, no API keys. Replaces LangExtract.

LLM-assisted BIO annotation and CRF sequence-labeling demo for Spanish pharmaceutical product names: builds a token-level corpus, trains a CRF model, tags new lines, and extracts structured fields.

Collection of purpose-built MCP servers for AI agent workflows.

Accounts-payable agent: LLM extraction of messy invoices + deterministic 3-way match (invoice/PO/receipt) + human-in-the-loop review, with a safety gate that no bad invoice is ever auto-approved. Fully offline.

Send your low-confidence document extractions. A human reviews them against the PDF and returns a typed Pydantic/Zod response. Managed document verification for AI agents. PDF + handwritten OCR. Client-side fragmentation: full document never leaves your machine. $0.80/page + $5 free credit. Express 30-min SLA. Built on open source awaithumans.

A simple llm library

news-summizr extracts structured summaries from headlines, labeling key points like announcement, products, region for quick insight.

A new package is designed to facilitate structured, reliable extraction of key insights from user-provided texts about cultural topics. It accepts a text input, such as an article or discussion prompt

Deterministic structured extraction from noisy LLM/OCR output. Zero LLM round-trips, microsecond latency, confidence score on every result. msgspec · Pydantic · dataclasses.

Structured CV extraction with strict JSON schema and anti-hallucination guarantees.

Turn tutorial videos into structured specs — Pine Script, recipes, code walkthroughs

Automated research paper analysis: PDF → JSON with evidence extraction using LLMs (DeepSeek, Gemma). Extracts methods, results, datasets, and claims with precise evidence grounding.

简历 PDF → 结构化数据 + 可读 HTML;电子版直取文字层、扫描件自动 OCR,多模态 LLM 读图 + source-span 规则校验防幻觉

schema-driven evaluation for LLM JSON extraction, json evaluation, structured-extraction, benchmark