Skip to content

LightRAG

v1.5.1 Breaking

This release includes 1 breaking change for platform teams planning a safe upgrade.

Published 1mo RAG & Retrieval
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

genai llm gpt-4 graphrag knowledge-graph large-language-models
+1 more
retrieval-augmented-generation

Affected surfaces

auth

ReleasePort's take

Light signal
editorial:auto 1mo

This release adds a Stop button for query cancellation and fixes several bugs across UI, parsing, and PostgreSQL validation.

Why it matters: Query cancellation improves user control; bugfixes ensure accurate error reporting, stable chunker processing, and correct Postgres search‑path handling in table checks.

Summary

AI summary

Broad release touches What's New, deps, chunker, and query.

Changes in this release

Feature Medium

Add query cancellation with Stop button.

Add query cancellation with Stop button.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Dependency Low

Update Redis requirement to >=5.0.0,<9.0.0.

Update Redis requirement to >=5.0.0,<9.0.0.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Dependency Low

Bump lucide‑react from 1.16.0 to 1.17.0 in web UI.

Bump lucide‑react from 1.16.0 to 1.17.0 in web UI.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Dependency Low

Bump katex from 0.16.47 to 0.17.0 in web UI.

Bump katex from 0.16.47 to 0.17.0 in web UI.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Bugfix Medium

Honor Postgres search path in table checks.

Honor Postgres search path in table checks.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Bugfix Medium

Surface accurate error for non‑ZIP .docx files in parser.

Surface accurate error for non‑ZIP .docx files in parser.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Bugfix Medium

Glue body‑less heading blocks into deeper child for chunker (P strategy).

Glue body‑less heading blocks into deeper child for chunker (P strategy).

Source: llm_adapter@2026-06-09

Confidence: high

 —
Bugfix Medium

Prevent double‑click on Send from aborting the query in web UI.

Prevent double‑click on Send from aborting the query in web UI.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Refactor Low

Remove few‑shot examples from the official entity‑relation extraction prompt.

Remove few‑shot examples from the official entity‑relation extraction prompt.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Refactor Low

Rename MAX_ASYNC to MAX_ASYNC_LLM with alias fallback in config.

Rename MAX_ASYNC to MAX_ASYNC_LLM with alias fallback in config.

Source: llm_adapter@2026-06-09

Confidence: high

 —
Full changelog

Importance Notes

To prevent weaker LLMs from inadvertently incorporating few-shot examples from the prompt into the extraction results, the official entity-relation extraction prompt has been updated in the current version by removing the few-shot examples. Empirical evaluation on the Qwen3-30B model shows no degradation in extraction quality. However, users are advised to monitor performance carefully on their specific LLMs to ensure consistent results.

To modify the system's default behavior, provide entity types and example prompt texts via env var ENTITY_TYPE_PROMPT_FILE.

What's New

  • feat(query): optionally send chunk heading path to the LLM by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3211
  • ✨ feat(chunker): inject split-table headers at split time, freeze split slices by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3218
  • feat(extract): inject section heading breadcrumb into extraction prompts by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3225
  • Add query cancellation with Stop button by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3227
  • ♻️ refactor: make extraction prompt examples format-only by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3231

What's Changed

  • build(deps): update redis requirement from <8.0.0,>=5.0.0 to >=5.0.0,<9.0.0 by @dependabot[bot] in https://github.com/HKUDS/LightRAG/pull/3194
  • 🔧 chore(mineru): default MINERU_LOCAL_IMAGE_ANALYSIS to false by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3208
  • build(deps): bump the react group in /lightrag_webui with 3 updates by @dependabot[bot] in https://github.com/HKUDS/LightRAG/pull/3199
  • build(deps): bump lucide-react from 1.16.0 to 1.17.0 in /lightrag_webui in the ui-components group by @dependabot[bot] in https://github.com/HKUDS/LightRAG/pull/3200
  • build(deps-dev): bump the build-tools group in /lightrag_webui with 3 updates by @dependabot[bot] in https://github.com/HKUDS/LightRAG/pull/3201
  • build(deps): bump katex from 0.16.47 to 0.17.0 in /lightrag_webui in the content-rendering group across 1 directory by @dependabot[bot] in https://github.com/HKUDS/LightRAG/pull/3202
  • build(deps): bump the frontend-minor-patch group across 1 directory with 2 updates by @dependabot[bot] in https://github.com/HKUDS/LightRAG/pull/3203
  • ♻️ refactor(config): rename MAX_ASYNC to MAX_ASYNC_LLM with alias fallback by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3209
  • fix: honor Postgres search path in table checks by @he-yufeng in https://github.com/HKUDS/LightRAG/pull/3206
  • fix(parser): surface accurate error for non-ZIP .docx files by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3210
  • ✨ fix(chunker): glue body-less heading blocks into their deeper child (P strategy) by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3214
  • ✨ fix(chunker): land parent-path gate + bridge overlap for P strategy by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3215
  • fix(mineru): preserve HTML table content by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3219
  • ✨ feat(multimodal): declare table content format (html/json) in analysis prompt by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3221
  • Preserve HTML table header spans in split tables by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3222
  • feat(sidecar): add parent_headings to tables/drawings/equations sidecars by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3223
  • fix(extract): drop "in section unknown" noise from multimodal relation descriptions by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3224
  • feat(query): cap parent-heading length to match extraction breadcrumb by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3226
  • Rename RetrievalTesting to RetrievalView by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3229
  • fix(webui): prevent double-click on Send from aborting the query by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/3230

Full Changelog: https://github.com/HKUDS/LightRAG/compare/v1.5.0...v1.5.1

Breaking Changes

  • Refactor: rename config key MAX_ASYNC to MAX_ASYNC_LLM with alias fallback
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 LightRAG

Get notified when new releases ship.

Sign up free

About LightRAG

[EMNLP2025] "LightRAG: Simple and Fast Retrieval-Augmented Generation"

All releases →

Related context

Earlier breaking changes

  • v1.5.3 One-time automatic migration copies legacy collection to model‑suffixed schema on upgrade.
  • v1.5.0rc3 Removes deprecated `QueryParam.model_func` field from lightrag/base.py.
  • v1.5.0rc3 Deprecates the `ENTITY_TYPES` environment variable; replace with ENTITY_TYPE_PROMPT_FILE.
  • v1.5.0rc2 ENTITY_TYPES environment variable deprecated; replace with ENTITY_TYPE_PROMPT_FILE before upgrading.

Beta — feedback welcome: [email protected]