---
type: "post"
id: "parse-extract-classify-now-each-with-its-own-9c14cb"
slug: "parse-extract-classify-now-each-with-its-own-9c14cb"
title: "Parse, Extract, Classify — now each with its own MCP"
description: "LlamaIndex’s July 15 update is a workflow-friction release: it moves schema creation toward conversation, makes a faster parsing path Markdown-first, and expands the default collaboration envelope."
retrieval_nugget: "LlamaIndex’s July 15 update is a workflow-friction release: it moves schema creation toward conversation, makes a faster parsing path Markdown-first, and expands the default collaboration envelope."
published_at: "2026-09-25"
updated_at: "2026-09-25"
record_date: "2026-09-25"
date_kind: "published_at"
topics: []
source_urls: ["https://www.llamaindex.ai/blog/llamaindex-newsletter-26-07-15"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:parse-extract-classify-now-each-with-its-own-9c14cb"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:0739d43f89d2a42c13f29ee846ceeeb0f9cb4c018aae0158b28f26e4e9b5a4c9","reviewed_at":"2026-09-25T20:05:00.485Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":2,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"LlamaIndex’s July 15 update is a workflow-friction release: it moves schema creation toward conversation, makes a faster parsing path Markdown-first, and expands the default collaboration envelope.","observed_facts":[{"text":"The source says Conversational Extract lets users generate an extraction schema through conversation rather than hand-writing JSON Schema.","source_urls":["https://www.llamaindex.ai/blog/llamaindex-newsletter-26-07-15"]},{"text":"The source says the fast tier returns clean Markdown by default with `tier=\"fast\", version=\"latest\"`, using LiteParse under the hood.","source_urls":["https://www.llamaindex.ai/blog/llamaindex-newsletter-26-07-15"]},{"text":"The source says every plan, including free, now has 100 users.","source_urls":["https://www.llamaindex.ai/blog/llamaindex-newsletter-26-07-15"]}],"mechanism":"The mechanism is that conversational schema design lowers the translation burden from a business request into a structured extraction specification, while standardized Markdown output gives downstream reviewers and agent workflows a more directly usable intermediate representation.","why_now":"Why now is that document-agent deployments increasingly fail on setup and review throughput rather than on a single model call; widening free-plan seats makes collaborative evaluation easier to begin.","implications":["The first implication is that teams can test extraction designs with more non-engineering participants before committing to a custom schema or paid collaboration setup.","Markdown-default output may simplify inspection and downstream text-oriented workflows, while still requiring document-specific validation."],"evidence_boundary":"The source is a vendor newsletter, so it reports product availability and intended behavior but supplies no benchmark, error-rate comparison, or independent evidence of production impact.","watch_conditions":["The first watch condition is whether LlamaIndex publishes accuracy, latency, or migration evidence comparing the new fast-tier Markdown path with prior parsing modes.","Whether Conversational Extract produces stable schemas across repeated, ambiguous business prompts."],"related_records":[],"new_branch_reason":"No prior New Runtime record is supplied with usable editorial content; this source establishes a distinct analytical record for the bundled schema, parsing-output, and collaboration changes."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/parse-extract-classify-now-each-with-its-own-9c14cb/","markdown":"https://newruntime.com/posts/parse-extract-classify-now-each-with-its-own-9c14cb.md","json":"https://newruntime.com/posts/parse-extract-classify-now-each-with-its-own-9c14cb.json"}
---

# Parse, Extract, Classify — now each with its own MCP

## Retrieval answer

LlamaIndex’s July 15 update is a workflow-friction release: it moves schema creation toward conversation, makes a faster parsing path Markdown-first, and expands the default collaboration envelope.

LlamaIndex’s July 15 update is a workflow-friction release: it moves schema creation toward conversation, makes a faster parsing path Markdown-first, and expands the default collaboration envelope. The practical significance is not a claimed breakthrough in document-understanding accuracy, but a lower-cost path to trying, sharing, and revising document-extraction workflows.

The source says Conversational Extract lets users generate an extraction schema through conversation rather than hand-writing JSON Schema. It also says the fast tier returns clean Markdown by default with `tier="fast", version="latest"`, using LiteParse under the hood, and that every plan—including free—now has 100 users. These changes combine to reduce handoff friction between the person defining fields, the developer configuring the pipeline, and teammates reviewing output.

The mechanism is that conversational schema design lowers the translation burden from a business request into a structured extraction specification, while standardized Markdown output gives downstream reviewers and agent workflows a more directly usable intermediate representation. Why now is that document-agent deployments increasingly fail on setup and review throughput rather than on a single model call; widening free-plan seats makes collaborative evaluation easier to begin. The source is a vendor newsletter, so it reports product availability and intended behavior but supplies no benchmark, error-rate comparison, or independent evidence of production impact.

The first implication is that teams can test extraction designs with more non-engineering participants before committing to a custom schema or paid collaboration setup. A second implication is that Markdown-default output may simplify inspection and downstream text-oriented workflows, while still requiring document-specific validation. The first watch condition is whether LlamaIndex publishes accuracy, latency, or migration evidence comparing the new fast-tier Markdown path with prior parsing modes. Also watch whether Conversational Extract produces stable schemas across repeated, ambiguous business prompts.

Source: https://www.llamaindex.ai/blog/llamaindex-newsletter-26-07-15

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)