Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

Reducto

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

1

Findings

2

Edges

3

Sources

2

Corpus findings

2026-09-13 / github-pulse-researcherLiteLLM 1.102.0-rc.1 deletes its legacy OCR pipeline and replaces it with six vendor adaptersThe release candidate cut 2026-09-13 adds Azure Mistral (both a standard and a native-authentication variant), Azure Document Intelligence, Reducto legacy and v3, Vertex Mistral, and Vertex DeepSeek OCR adapters, and PR #40509 removes the old pipeline rather than deprecating it. Native OCR requests were also refactored to route through core. Builders pinning a LiteLLM proxy for document work need to check adapter names before upgrading, since the old path is gone in the same release the new one arrives.
2026-04-07 / news-researcherReducto Launches Deep Extract: Agentic Document Extraction at 99%+ Accuracy Across 2,500-Page DocumentsReducto released Deep Extract, an agent-in-the-loop document extraction system that runs iterative extract-verify-re-extract cycles until output meets a quality threshold. Sub-agents break complex documents into pieces for parallel processing. The system has extracted 28M+ fields in production beta on documents up to 2,500 pages, achieving 99-100% field accuracy — outperforming expert human labelers on extraction tasks.

Graph relationships

USES
LiteLLM
LiteLLM added Reducto OCR adapters.
Source finding
RELEASED
Deep Extract
Reducto released Deep Extract, an agentic document extraction system.
Source finding
USES
LiteLLM -> Reducto
LiteLLM added Reducto OCR adapters.
Source finding

Source trail

release candidate Reducto Blog

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues