Skip to content

langroid

v0.65.0 Breaking

This release includes 2 breaking changes for platform teams planning a safe upgrade.

Published 6d AI Agents & Assistants
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agents ai chatgpt function-calling llm gpt-4
+10 more
gpt4 information-retrieval language-model llama llm-agent llm-framework local-llm multi-agent-systems openai-api retrieval-augmented-generation

Affected surfaces

breaking_upgrade

ReleasePort's take

Moderate signal
editorial:auto 6d

The default PDF parser now uses pypdfium2 instead of pymupdf4llm, removing the AGPL dependency. The pymupdf packages are removed from core and offered as optional extras.

Why it matters: Affects any code relying on pymupdf4llm or pymupdf for PDF parsing; migration required before next upgrade cycle.

Summary

AI summary

Updates What changed, Breaking change & migration, and https://github.com/py-pdf/benchmarks across a mixed release.

Changes in this release

Breaking High

Default PDF parser switched from pymupdf4llm to pypdfium2; AGPL dependency removed.

Default PDF parser switched from pymupdf4llm to pypdfium2; AGPL dependency removed.

Source: llm_adapter@2026-05-29

Confidence: high

 —
Feature Low

`pypdfium2` added as core dependency; `PyPDFium2Parser` extracts text page‑by‑page.

`pypdfium2` added as core dependency; `PyPDFium2Parser` extracts text page‑by‑page.

Source: llm_adapter@2026-05-29

Confidence: high

 —
Deprecation Medium

`pymupdf4llm` and `pymupdf` removed from core dependencies; available as opt‑in extras.

`pymupdf4llm` and `pymupdf` removed from core dependencies; available as opt‑in extras.

Source: llm_adapter@2026-05-29

Confidence: high

 —
Full changelog

Permissive-by-default PDF parsing — no AGPL in the default install

Langroid is MIT-licensed, but until now a plain pip install langroid pulled in
pymupdf4llm (and transitively pymupdf), which are AGPL-3.0 licensed. This
release removes that AGPL dependency from the default install and switches the
default PDF parser to the permissively-licensed pypdfium2
(Apache-2.0 / BSD-3-Clause). Resolves #1026.

What changed

  • pypdfium2 is now the default PDF parser, added as a core dependency. A new
    PyPDFium2Parser extracts text page-by-page via PDFium.
  • pymupdf4llm / pymupdf removed from core dependencies. They remain
    available as opt-in extras: doc-chat, pdf-parsers, all, or pymupdf4llm.
  • DocChatAgent now defaults to pypdfium2 as well, so document-chat works
    out of the box on a bare install with no AGPL code.
  • Per the py-pdf benchmarks, pypdfium2
    matches or exceeds pymupdf on raw text-extraction accuracy.

Breaking change & migration

  • A bare pip install langroid no longer installs pymupdf4llm/pymupdf, and the
    default PDF parser now emits plain text rather than pymupdf4llm's structured
    Markdown.

  • If you want the richer Markdown extraction (headers, tables, multi-column reflow)
    from pymupdf4llm, install an extra and select it explicitly:

    pip install "langroid[doc-chat]"   # or [pdf-parsers], [all], [pymupdf4llm]

    from langroid.parsing.parser import ParsingConfig, PdfParsingConfig

cfg = ParsingConfig(pdf=PdfParsingConfig(library="pymupdf4llm"))

Thanks to @alexagr for reporting the licensing issue (#1026). See #1028 for details.

Breaking Changes

  • Removed core dependencies `pymupdf4llm` and `pymupdf`; they are now opt‑in extras (doc-chat, pdf-parsers, all, pymupdf4llm).
  • Default PDF parser changed from `pymupdf4llm` to `pypdfium2`, altering output from structured Markdown to plain text.
View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track langroid

Get notified when new releases ship.

Sign up free

About langroid

Harness LLMs with Multi-Agent Programming

All releases →

Related context

Earlier breaking changes

  • v0.64.0 Rejects PostgreSQL/SQLite/MSSQL primitives that enable arbitrary file reads.
  • v0.64.0 Restricts file‑tool paths to stay within configured `curr_dir`.

Beta — feedback welcome: [email protected]