Explore 2387 PostgreSQL extensions from PGEXT, maintained by Pigsty.
134 extensions found. Clear filters
| Name | Version | Description | Category | License | Language | PG Version | Stars |
|---|---|---|---|---|---|---|---|
| llama_embed | 0.1.0 | llama.cpp embedding extension for dense and multivector search | RAG | PostgreSQL | C | — | 9 |
| pg_colbert_llama | 0.1.0 | ColBERT embedding companion extension for pgturbohybrid | RAG | PostgreSQL | C | — | 9 |
| pgturbohybrid | 0.2.0 | hybrid vector and BM25 search for PostgreSQL on top of pgvector | RAG | PostgreSQL | C | — | 9 |
| pgturbohybrid_experimental | 0.2.0 | experimental sparse and multivector features for pgturbohybrid | RAG | PostgreSQL | C | — | 9 |
| maludb_core | 0.104.0 | A memory DBMS for long-term institutional memory, human-AI knowledge sharing, and contextual recall. PostgreSQL 17 extension. | RAG | — | C | — | 8 |
| pg_semantic_cache | 0.1.0-beta4 | Caches query results and retrieves semantically similar results using pgvector embeddings and configurable eviction policies. | RAG | PostgreSQL | C | 14–18 | 8 |
| sveddy | 0.1.0 | A PostgreSQL extension for collaborative filtering using low-rank matrix approximation techniques. | RAG | MIT | C | — | 8 |
| bktree | 1.0 | SP-GiST BK-tree index extension for 64-bit perceptual hashes and Hamming-distance searches. | RAG | — | C | 15 | 7 |
| hamming_distance | 0.0.1 | Hamming distance and similarity functions for hexadecimal feature-vector strings. | RAG | MIT | C | — | 7 |
| pg_cuvs | 0.5.0 | ⚡ Blazing-fast GPU vector search for PostgreSQL powered by NVIDIA cuVS. | RAG | PostgreSQL | C | — | 7 |
| quria | 0.0.0 | quria: experimental pgrx extension for BM25 full-text and HNSW vector search | RAG | Apache-2.0 | Rust | 11–15 | 7 |
| pg_summarize | 0.0.1 | Text Summarization using LLMs. Built using pgrx | RAG | — | Rust | 14–18 | 6 |
| pgmnemo | 0.20.0 | Single-plan multimodal agent memory for PostgreSQL | RAG | Apache-2.0 | SQL | 17–18 | 6 |
| rikai | 0.1 | rikai ML extensions | RAG | Apache-2.0 | SQL | — | 5 |
| pg_ask | 1.0 | Postgres Extension which uses AI to create SQL from Natural language. | RAG | MIT | C++ | 16–18 | 4 |
| pg_rvbbit | 4.1.4 | Semantic SQL operators backed by configurable LLMs, with caching, receipts, and routing across multiple execution engines. | RAG | — | Rust | — | 4 |
| pg_vector_embedding | 1.0.0 | Trigger-driven vector embedding through external HTTP services and a background queue. | RAG | Apache-2.0 | SQL | — | 4 |
| tlsh_gist | 1.0 | TLSH fuzzy-hash data type, distance operations, and GiST indexing for similarity search. | RAG | — | C++ | — | 4 |
| ecaz | 0.1.1 | Rust PostgreSQL extension for high-performance vector storage with broad quantization and index support. | RAG | MIT | Rust | — | 3 |
| mtree | 0.0.1 | Metric-space index access method derived from PostgreSQL GiST | RAG | — | C | — | 3 |
| otlet | 0.1.0 | Local LLM inference inside PostgreSQL with background workers, queued tasks, receipts, and semantic data operations. | RAG | — | Rust | — | 3 |
| pg_extension | 0.1.0 | In-database model slicing and inference helper functions for structured data analytics. | RAG | — | Rust | 11–15 | 3 |
| pg_infer | 1.0.0 | Expose transformer model knowledge as SQL relations and a custom index access method | RAG | Apache-2.0 | Rust | 18 | 3 |
| pg_rrf | 0.0.3 | Reciprocal rank fusion functions for hybrid search | RAG | MIT | Rust | 14–18 | 3 |
| pg_tiktoken_c | 1.1 | Fast tiktoken BPE tokenizer for PostgreSQL implemented in C | RAG | Apache-2.0 | C | 14–18 | 3 |
| aqua | 0.0.1 | In-database neural user-defined functions for machine-learning inference. | RAG | — | C | — | 2 |
| flame | 0.0.1 | PGRX proof-of-concept extension that generates text embeddings with fastembed/MiniLM. | RAG | AGPL-3.0 | Rust | 11–16 | 2 |
| pg_knowledge_graph | 0.1.0 | Knowledge graph extension for PostgreSQL with graph algorithms and pgvector integration. | RAG | MIT | Rust | 16–18 | 2 |
| pgembed | 0.1.0 | Generate vector embeddings through Ollama, OpenAI, or custom HTTP services | RAG | LicenseRef-Tecnisys-Community-1.0 | Python | 13–18 | 2 |
| pgtoken | 1.0 | PostgreSQL extension for compact, codebook-compressed LLM token ID storage. Encode once, retrieve forever. O(1) token count without decoding. | RAG | MIT | Python | — | 2 |
| postllm | 0.1.0 | postllm: PostgreSQL-native access to OpenAI-compatible and local LLM runtimes | RAG | MIT | Rust | 13–18 | 2 |
| sparse_vector | 0.1 | C data type and functions for sparse float vectors, normalization, dot product, and cosine similarity. | RAG | MIT | C | 10–18 | 2 |
| svs | 0.1.0 | SVS Vamana approximate-nearest-neighbor index with LeanVec and LVQ compression for pgvector. | RAG | PostgreSQL | C | — | 2 |
| ai_toolkit | 1.0 | AI-powered query generation and explanation tools for PostgreSQL. | RAG | — | C++ | 18 | 1 |
| hamming | 0.1.0 | Rust/pgrx extension providing a Hamming-distance operator for bytea plus an unfinished experimental IVF index access method. | RAG | — | Rust | 11–16 | 1 |
| hclust | 0.0.1 | C function for agglomerative or divisive hierarchical clustering that writes cluster assignments back to a table. | RAG | — | C | — | 1 |
| lshvector | 1.0 | a feature vector type and a locality sensitive hashing index | RAG | — | Python | — | 1 |
| malwaredb_pg | 0.0.1-dev | PostgreSQL extension for MalwareDB similarity functions: SSDeep, LZJD, and TLSH | RAG | Apache-2.0 | Rust | 13–18 | 1 |
| paves | 0.1 | Filtered L2 vector search with CPU and optional GPU execution | RAG | — | C | 14 | 1 |
| pg_fela | 1.0.0 | The holy grail of AutoML for PG - self classifying and regressing tabular data! | RAG | — | Rust | — | 1 |
| pg_gen_query | 1.0 | A PostgreSQL extension written in C++ that adds no-maintenance GenAI querying functionality using Clickhouse's AI SDK | RAG | MIT | C++ | — | 1 |
| pg_grammar_guard | 0.4.1 | Catalog-derived grammars and approved grammar drift checks | RAG | PostgreSQL | SQL | 14–18 | 1 |
| pg_ocr | 0.0.1 | Self-contained OCR functions via PL/Python | RAG | — | Python | — | 1 |
| pg_optigraph | 0.1.0 | pg Optigraph is a PostgreSQL extension which uses OptiGraph ML model. | RAG | MIT | Python | — | 1 |
| pg_pageindex | 1.0 | PageIndex document-tree construction and tree search for PDF and Markdown through a C/Go bridge. | RAG | MIT | C | 16 | 1 |
| pg_recall_guard | 0.2.0 | Measure ANN index recall and detect drift from an approved baseline | RAG | PostgreSQL | PLpgSQL | 13–19 | 1 |
| pg_splade | 0.0.0 | Generate SPLADE sparse-vector embeddings inside PostgreSQL | RAG | — | Rust | 12–17 | 1 |
| pg_synapse_pgrx | 0.1.1 | pg synapse: Postgres-native agent-loop runtime | RAG | — | Rust | — | 1 |
| pg_turbovec | 1.29.0 | TurboQuant-compressed vector type and ANN index access method for PostgreSQL. | RAG | Apache-2.0 | Rust | 14–18 | 1 |
| pgokf | 0.1.13 | Transactional Open Knowledge Format catalog with lexical and optional semantic search. | RAG | AGPL-3.0-only | Rust | 15–19 | 1 |