{"full_name":"HKUDS/LightRAG","owner":"HKUDS","name":"LightRAG","shard":"1","url":"https://github.com/HKUDS/LightRAG","description":"[EMNLP2025] LightRAG: Simple and Fast Retrieval-Augmented Generation","homepage":"https://arxiv.org/abs/2410.05779","topics":["knowledge-graph","large-language-models","retrieval-augmented-generation","genai","graphrag","llm","rag","gpt","docling","mineru","ragas"],"tag_pages":["genai","gpt","graphrag","knowledge-graph","large-language-models","llm","rag","retrieval-augmented-generation"],"language":"Python","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/118165258?v=4","created_at":"2024-10-02T11:57:54Z","pushed_at":"2026-11-15T17:12:42Z","metrics":{"stars":40707,"forks":5779,"watchers":287,"issues":1684,"prs":2582,"releases":84,"issues_open":232,"issues_closed":1452,"prs_open":164,"prs_closed":499,"prs_merged":1919,"discussions":133},"growth":{"day":{"stars":14,"forks":3,"base_day":"2026-11-15"},"week":{"stars":86,"forks":10,"base_day":"2026-11-09"}},"history":{"dates":["2026-11-09","2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[40621,40636,40659,40673,40673,40687,40693,40707],"forks":[5769,5770,5773,5773,5774,5775,5776,5779],"watchers":[285,285,286,286,286,286,286,287],"issues_open":[null,null,244,243,241,235,235,232],"prs_open":[null,null,190,185,180,170,167,164]},"boards":{},"same_owner":[{"full_name":"HKUDS/AI-Researcher","name":"AI-Researcher","shard":"1","boards":{}},{"full_name":"HKUDS/AI-Trader","name":"AI-Trader","shard":"1","boards":{}},{"full_name":"HKUDS/AutoAgent","name":"AutoAgent","shard":"1","boards":{}},{"full_name":"HKUDS/AutoSwarm","name":"AutoSwarm","shard":"1","boards":{}},{"full_name":"HKUDS/CLI-Anything","name":"CLI-Anything","shard":"1","boards":{"monthly":87}},{"full_name":"HKUDS/ClawTeam","name":"ClawTeam","shard":"1","boards":{}},{"full_name":"HKUDS/DeepCode","name":"DeepCode","shard":"1","boards":{}},{"full_name":"HKUDS/DeepTutor","name":"DeepTutor","shard":"1","boards":{"daily":87,"weekly":66,"monthly":24,"monthly_rate":79,"daily_forks":190}},{"full_name":"HKUDS/OpenHarness","name":"OpenHarness","shard":"1","boards":{}},{"full_name":"HKUDS/Vibe-Trading","name":"Vibe-Trading","shard":"1","boards":{"daily":188,"monthly":185,"daily_watchers":78}},{"full_name":"HKUDS/nanobot","name":"nanobot","shard":"1","boards":{}}],"issues":[{"number":1323,"title":"[Feature Request]: Automatic merging of the same entity under different names","comments":35,"created_at":"2025-04-09T13:24:31Z"},{"number":2133,"title":"[Question]:How to implement multi tenant solutions?","comments":14,"created_at":"2025-09-22T11:53:13Z"},{"number":3631,"title":"[Design Review]: Server-side Multi-Workspace Phase 1","comments":13,"created_at":"2026-08-11T17:08:53Z"},{"number":1985,"title":"[Question]:How to add custom metadata columns for document management?","comments":13,"created_at":"2025-08-20T05:23:48Z"},{"number":3664,"title":"RFC: Add C strategy for Server-injected custom chunking callbacks","comments":12,"created_at":"2026-08-18T03:32:25Z"},{"number":2285,"title":"Can we temporarily enable/disable documents in RAG?","comments":11,"created_at":"2025-10-30T06:44:44Z"},{"number":2138,"title":"[Bug]: 🐛 Bug Report: QueryParam ids parameter not working in API Server Describe the bug","comments":11,"created_at":"2025-09-23T17:43:37Z"},{"number":2642,"title":"Guidance on Adding Multimodal Support to LightRAG: Wrap with RAG‑Anything or Extend (modify) LightRAGs lightrag‑server?","comments":10,"created_at":"2026-01-27T23:32:32Z"}],"discussions":[{"number":4403,"title":"Chopper: a small open model that cuts only the needed sentences out of retrieved text","comments":0,"category":"Show and tell","created_at":"2026-11-14T09:07:54Z"},{"number":4301,"title":"v1.5.6: supported chunk-selection diagnostics and applicability of #3887/#3904","comments":2,"category":"Q&A","created_at":"2026-11-05T02:30:14Z"},{"number":4150,"title":"Putting a dollar cap on one LightRAG indexing job","comments":0,"category":"Show and tell","created_at":"2026-10-03T16:45:26Z"},{"number":3933,"title":"Replaying a LightRAG run offline: exact with a local embedder, and exactly where it stops with openai_embed","comments":0,"category":"Show and tell","created_at":"2026-09-13T17:13:38Z"},{"number":3840,"title":"We wired LightRAG into a full-duplex voice runtime — ask your knowledge base questions by voice","comments":0,"category":"Show and tell","created_at":"2026-09-06T10:12:58Z"},{"number":3388,"title":"Feature Idea: Metadata filtering for retrieval (pre-filter chunks/entities by tags, file path, etc.)","comments":0,"category":"Ideas","created_at":"2026-07-10T14:06:58Z"},{"number":3342,"title":"LightRAG Upload Fails with \"NanoVectorDBStorage[entities] index flush failed: 404 page not found\"","comments":3,"category":"Q&A","created_at":"2026-06-28T09:58:15Z"},{"number":3341,"title":"File Upload Fails in LightRAG with \"llama.cpp terminated unexpectedly\" Error","comments":0,"category":"Q&A","created_at":"2026-06-28T09:27:06Z"}],"latest_release":{"name":"v1.5.8rc1","tag":"v1.5.8rc1","published_at":"2026-11-15T13:03:37Z","body":"## Important Notes\r\n\r\n> v1.5.8rc1 is a release candidate. Please test it on a copy of your data before upgrading production.\r\n\r\n### Highlights\r\n\r\n**1. Configuration storage: LightRAG now remembers which embedding model built your vectors**\r\n\r\nIn earlier releases, changing `EMBEDDING_MODEL` or `EMBEDDING_DIM` on an existing deployment failed silently. Depending on the backend, LightRAG either kept the old vectors and returned wrong neighbours, or opened a new, empty vector container (Milvus / Qdrant / PostgreSQL) and served a knowledge graph with no vectors at all.\r\n\r\nv1.5.8 adds a **configuration storage**. It is a small KV namespace (`config`) that holds each workspace's **embedding baselines**: for every vector target (entities, relationships, chunks), the embedding space its vectors were written in. Startup checks the baselines against the current embedding configuration. If the space has changed, the server **refuses to start** and names the target and the way out, instead of serving bad results.\r\n\r\n* **Backends:** `JsonKVStorage`, `PGKVStorage`, `MongoKVStorage` and `OpenSearchKVStorage`, selected with `LIGHTRAG_CONFIG_STORAGE`. Left unset, it follows `LIGHTRAG_KV_STORAGE`, so most deployments need no new setting. `RedisKVStorage` is not admitted, so Redis deployments default to `JsonKVStorage`, with a startup warning.\r\n* **No preparation needed.** On the first start of v1.5.8, an existing deployment records its baselines on its own. A baseline is recorded only on evidence: an **adoption probe** re-embeds a few stored records (up to 8 per target) and compares them with the stored vectors, or the container is confirmed empty. A probe that cannot run (for example, the embedding service is down) records nothing and is retried on the next start.\r\n* **Deployment anchor.** The first start also binds the deployment to the configuration container it opened, in `WORKING_DIR/config_storage_anchor.json` (backend type + container UUID). Every later start refuses a container of another type or another UUID. A changed `LIGHTRAG_KV_STORAGE`, a different database or a replaced volume therefore no longer silently moves your configuration.\r\n* **New offline tools:**\r\n  * `lightrag-rebuild-vdb` re-embeds a workspace with the current model and records the new baselines. This is the way out after an intentional model change.\r\n  * `lightrag-migrate-config` moves the configuration storage to a backend of another type, for example from JSON files to PostgreSQL.\r\n  * `lightrag-clear-storage` clears a workspace offline, configuration records included.\r\n* **Several servers, one `WORKING_DIR`:** supported when each server has its own `WORKSPACE` and port. A second server on the **same** workspace is refused where the filesystem supports file locks.\r\n\r\nFor more information, see:\r\n\r\n- [`docs/LightRAG-API-Server.md`](https://github.com/HKUDS/LightRAG/blob/main/docs/LightRAG-API-Server.md): *Configuration Storage Migration Between Backend Types*\r\n- [`lightrag/tools/README_REBUILD_VDB.md`](https://github.com/HKUDS/LightRAG/blob/main/lightrag/tools/README_REBUILD_VDB.md), [`README_MIGRATE_CONFIG.md`](https://github.com/HKUDS/LightRAG/blob/main/lightrag/tools/README_MIGRATE_CONFIG.md), [`README_CLEAR_STORAGE.md`](https://github.com/HKUDS/LightRAG/blob/main/lightrag/tools/README_CLEAR_STORAGE.md)\r\n- [`docs/design/ConfigurationStorageDesign.md`](https://github.com/HKUDS/LightRAG/blob/main/docs/design/ConfigurationStorageDesign.md): the design overview\r\n\r\n**2. Safer graph writes and document deletion**\r\n\r\n* Manual graph edits (create / edit / delete / merge entities and relations, `insert_custom_kg`) are now serialized against each other and against the ingestion pipeline on `NetworkXStorage`.\r\n* Commit ordering for entity/relation create, edit, merge and rename was reworked to prevent data loss when an operation is interrupted.\r\n* Interrupted document deletions are visible and retryable.\r\n* New `lightrag-repair-chunk-tracking` tool removes orphan chunk-tracki","is_prerelease":false},"readme":true,"readme_zh":true,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}