AI & machine learning

Hierarchical pentagonial indexing for geospatial data that is equal-area and millimeter accurate.

Connects DuckDB to any database that has an Arrow Database Connectivity (ADBC) driver.

An ADBC client extension for DuckDB that can access ADBC provided data sources.

AI functions for SQL — completions, classification, extraction, embeddings, and read-only SQL generation across local and hosted model providers

The Airport extension brings Arrow Flight support to DuckDB, enabling DuckDB to query, modify, and store data from Arrow Flight servers.

Efficiently query large chess game collections

Read AnnData (.h5ad) files for single-cell genomics data analysis, with support for local and remote (HTTP/HTTPS/S3) files

Statistical timeseries forecasting in DuckDB. Support ARIMA, SARIMA, ETS, TBATS, MFLES, MSTL, and other models.

Combinatorial decision algorithms as SQL — bin packing, knapsack, wave batching, sequencing, scheduling, assortment and portfolio selection

Git-like branching for analytical databases. Attach isolated what-if scenarios as catalogs, edit them with ordinary SQL on copy-on-write delta storage, branch, diff, and merge them back.

Multi-modal product similarity search for manufacturing supply chain planning. Find similar materials by BOM component overlap (Jaccard), graph-structural similarity (Weisfeiler-Lehman kernel), text embeddings, and transactional patterns, with predecessor detection and cold-start analog matching for demand forecasting.

A DuckDB extension for statistical regression and inference in SQL - OLS, Ridge, Elastic Net, LARS/LassoLars, WLS, recursive least squares, robust estimators (Huber, RANSAC, Theil-Sen), GLMs (Poisson, Negative Binomial, Binomial, Tweedie, Gamma, Logistic) with offset support, mixed-effects/hierarchical GLMs with random slopes and crossed/nested factors, AFT survival regression, explicit priors with Laplace intervals, empirical-Bayes shrinkage, and time-series regression, all with full diagnostics and inference.

Zero-shot tabular machine learning inside DuckDB — classification, regression, synthetic-data generation and imputation with real tabular foundation models (Mitra, TabPFN v2/2.5/3, TabICL, Orion-BiX, TabFM) on ONNX Runtime, no training loop

A duckdb extension which combines data quality and data preparation tools for tabular data.

Render charts and dashboards straight from SQL — the ggplot-rs grammar of graphics, emitted as SVG.

This extension is an alias to the nanoarrow extension. Allows the consumption and production of the Apache Arrow interprocess communication (IPC) format, both from files and directly from stream buffers.

Astronomical calculations + FITS file I/O - coordinate transforms, photometry, cosmology, dust extinction, sidereal time

Read Azure Blob / ADLS Gen2 (abfss/az/wasbs) over HTTPS+SAS, without the Azure SDK (Wasm-friendly)

Provides high-performance, space-efficient probabilistic data structures—including quotient, XOR, and binary fuse filters—for fast approximate set membership testing with no false negatives and configurable false positive rates.

Live SQL Queries on Blockchain

Secure remote secrets storage with OPAQUE PAKE authentication, email/password registration, MFA login, and automatic ducklake mounting

Get locally homebrew casks, packages and formulas and their dependencies as nicely types tables

BVH (BioVision Hierarchy) motion capture file parser for DuckDB

Read cached filesystem for httpfs

Prewarm data blocks into DuckDB's buffer pool or OS page cache for faster queries

Provides a hello world example demo from the C/C++ C API template

Astronomical coordinates utilities

Creates chaos! ⋆✴︎˚｡⋆ Chaos allows you to throw any type of DuckDB exception, or to raise a SIGSEGV, SIGABRT, or SIGBUS signal.

Provides mathematical clamp functions for numeric types in DuckDB, including general clamping, saturate, and others. Ensures values stay within a specified range and handles NULLs and NaNs correctly.

AWS CloudFront signed cookie authentication for DuckDB httpfs

Query SharePoint, OneDrive, Google Drive, Dropbox, SFTP, and any VPS (via cloudfs-agent) using spfs://, odfs://, gdfs://, dbxfs://, sftp://, and vfs:// — with the same capabilities as s3://. Supports Parquet, CSV, JSON, Delta Lake, Iceberg, glob patterns, and COPY TO across all providers.

Read and write Amazon CloudWatch telemetry with OTLP-shaped SQL tables

GDAL-free Cloud-Optimized GeoTIFF (COG) reader — query remote rasters in place over HTTP range reads, with STAC catalog integration

Cloud-Optimized ZIP reader for DuckDB

SQL-native web crawler with HTML extraction and MERGE support

DuckDB HTTP Cronjob Extension

Cryptographic hash functions and HMAC

httpfs with connection pool, HTTP/2 and async IO.

DuckDB filesystem extension for CWIQ FS

A fully visual data explorer and GUI for DuckDB. Create interactive dashboards and DAG pipelines.

Read logs from the Datadog Logs Search API into OTLP-shaped tables

By utilizing the Apache DataSketches library this extension can efficiently compute approximate distinct item counts and estimations of quantiles, while allowing the sketches to be serialized.

Read Databento DBN (Databento Binary Encoding) market-data files, including Zstd-compressed .dbn.zst captures, directly as SQL tables

Extension that adds support special decimal arithmetic functions

Lazy Machine Learning Imputation Architecture in DuckDB.

Attach a directory of Delta tables as a single DuckDB database with automatic schema discovery

Export DuckLake catalog tables to Delta Lake format

DuckDB integration with medical imaging data (DICOM - Digital Imaging and Communication in Medicine).

The DNS Extension enables DNS lookups and reverse DNS lookups from within DuckDB

Define and run data-quality tests inside DuckDB

Estimate Parquet scan bytes for SELECT queries without executing them

Read and write Stata .dta files (formats 117-121, Stata 13-18) with full type mapping, value labels, and strL support.

Build, transform, validate, and extract content from structured documents using the duck_block type

DuckDB Extension for Delta Sharing

DuckDB extension for discrete global grid systems (DGGS) powered by DGGRID v8.

Diff two relations (tables, queries) off a primary key, reporting per-key status (identical/different/left_only/right_only), a JSON summary of changed columns, and per-column left/right values.

DuckDB extension for converting GEOMETRY/WKB to and from GeoArrow native encodings, powered by geoarrow-c. Built against DuckDB v1.5.5

Parse and analyze test results, build outputs, and CI/CD pipeline logs from 110+ formats with severity filtering, format auto-detection, and context extraction

DuckLineage automatically captures data lineage from every DuckDB query and emits OpenLineage events.

Query LabKey Server tables with automatic local Parquet caching

Smart Development Intelligence for DuckDB - Git-aware data analysis capabilities that allow querying git history, accessing files at any revision, and performing version-aware data analysis with SQL.

The most efficient way to query Delta Lake tables directly from DuckDB via the Delta Sharing protocol.

Retrieve geo ip data from MMDB database

Model Context Protocol (MCP) extension for DuckDB that enables seamless integration between SQL databases and MCP servers. Provides both client capabilities for accessing remote MCP resources via SQL and server capabilities for exposing database content as MCP resources.

Remote storage filesystem for DuckDB via Apache OpenDAL — S3, GCS, Azure, Hugging Face, WebDAV, SFTP, and more, with DuckDB secret-based auth.

Cheminformatics with RDKit for DuckDB - SMILES/SDF molecules, substructure search, and molecular descriptors.

Phonetic algorithms for German and European name matching

Explore Zarr stores with SQL through relational metadata, chunk discovery, and dense cell scans

Browser-based BI dashboard extension for DuckDB. Run SQL queries and build interactive dashboards with drag-and-drop charts, all from the DuckDB CLI.

DuckDB geospatial visualization with deck.gl and MapLibre

Adds ISO GQL graph querying and graph algorithms to DuckDB

Stream changes from your DuckLake — durable consumer cursors, single-reader windows, typed DDL events.

Asset-centric data orchestration with JSON ingestion, partitions, declarative automation sensor, Snowflake-compatible CREATE DYNAMIC ASSET, MCP server, lineage and OpenLineage emission — all on DuckDB

Cheminformatics toolkit for DuckDB - SMILES, InChI, MOL/SDF, PDB, SELFIES, Wildman-Crippen LogP, Morgan/ECFP fingerprints, and Tanimoto similarity from SQL

Intelligent query result caching between DuckDB and Snowflake with TTL and smart refresh

ASK() — natural-language SQL for DuckDB, grounded in your dbt Semantic Layer

DuckDB C extension for in-process JIT compiled C UDFs via TinyCC — self-contained, no external runtime required

Extension that adds support for working with the EEA Reference Grid System.

The long-awaited Dutch localization of DuckDB's SQL dialect

Read & write SAP IDoc files (flat + IDoc-XML) as SQL in DuckDB — a community extension in the erpl family.

Reach any TCP service from DuckDB through an SSH bastion, Tailscale or NetBird — and publish local ports back onto those networks.

ERPL is a DuckDB extension to connect to API based ecosystems via standard interfaces like OData, GraphQL and REST. This works e.g. for SAP ERP, SAP Ariba, SAP C4C, Microsoft Dynamics and many more.

Extension that adds support for reading data from EUROSTAT database using SQL.

Evaluate the Rhai scripting language in DuckDB

A extension that hooks into database events and sends JSON-formatted notifications to external programs.

Provides access to faiss indices from DuckDB.

The FakeIt Extension enables generation of realistic fake/test data directly in SQL queries

Choose a file via native file dialog

SQL-native quant finance functions for DuckDB

Semantic type classification — detects 251 data types (emails, URLs, dates, UUIDs, currencies, etc.) from raw strings

Query Google Cloud Firestore directly from DuckDB using SQL

Federated read-only access to Firebird (3.0/4.0/5.0) databases from DuckDB, with projection + filter pushdown, INT128 / DECIMAL(38) / TIMESTAMP_TZ support, native ATTACH, and CHARACTER SET NONE handling (default win1252; strict / iso8859_1 / blob also available) with filter-pushdown correctness on transcoded columns.

Read Garmin .fit files using DuckDB - GPS tracks, heart rate, power metrics, and fitness device data

Pack sparse join build side into compact variants.

LLM & RAG extension to combine analytics and semantic analysis

Query filesystem metadata using SQL with recursive directory traversal

Dynamic function invocation - call any scalar function or macro by name at runtime

Query Oracle Fusion from DuckDB through BI Publisher, with SSO, an ATTACH-able catalog and cached metadata

Fuzzy matching based autocompletion

A DuckDB extension for working with Kaggle datasets

Query Google Cloud logs, metrics, alerts, and service topology from DuckDB

DuckDB GCS Extension

Query files in Google Drive directly with a gdrive:// filesystem

Compact geometry encoding using delta-encoded coordinates. Achieves 3-4x smaller geometry storage and wire transfer compared to WKB.

Bindings to ggsql — Grammar of Graphics based visualizations in SQL

Experimental gh:// filesystem support

Read GORpipe .gorz and .gord files and write .gorz as native DuckDB tables

GPU-accelerated aggregates and joins for DuckDB on NVIDIA CUDA and Apple Silicon Metal — upload columns once, then reductions and joins run on the GPU. First SQL execution engine that targets Apple Silicon GPUs.

Offline geocoding from a single country file — forward and reverse, no server, no network

Read and write Google Sheets using SQL

Read HDF5 datasets and attributes

Client/server DuckDB over one HTTP port — web UI, JSON /sql, and native ATTACH (TYPE harbor)

Non-cryptographic hash functions, xxHash, rapidhash and Murmurhash3

Read and write hdfs:// URLs via a native HDFS filesystem — no JVM or libhdfs required

Provides hedged request for filesystem operations

Hex9 hierarchical hexagonal DGGS — point encoding, layer binning, cell geometry as WKB, labels, k-rings, Hamiltonian-curve ordering, and grid enumeration

Attach Apache Hive Metastore as a native DuckDB catalog and read/write the data inside with ease!

Holt-backed metadata indexes for fast DuckDB file discovery

Navigate and explore the filesystem using SQL

DuckDB HTTP Client Extension

Better HTTP statistics for DuckDB

Add per-operation timeout and retry for httpfs extension

Discover, query, profile, and estimate the storage of Hugging Face datasets

A DuckDB extension for in-database inference

Powerful string case transformation and inflection capabilities directly to your SQL queries.

Native INTERLIS model introspection and streaming XTF support for DuckDB

Read AWS Ion data and write Ion text or binary with typed mappings

Adds support for validating JSON data with JSON schemas.

The JSONata extension for DuckDB enables expressive, JSON-focused querying and transformation directly within SQL using the powerful JSONata expression language.

Analytics-optimized JSON storage — parse one time and read many times, and shred hot paths into typed columns for pushdown and row-group pruning

DuckDB extension for Keboola Storage — query and write Keboola tables using standard SQL

Reader for the Lastra columnar time-series format

Used to inject simulated latency to filesystem operations

Read AGS4 geotechnical data files as typed, UUID-keyed tables directly from SQL — born-typed columns, deterministic content-addressed keys that join across groups by construction, and an embedded AGS dictionary. A read-only SQL surface. Local, http(s):// and s3:// (with httpfs).

Generate LDBC SNB BI datasets directly from DuckDB

Storage extension wrapping LevelDB with pivot semantics for structured relational tables over key-value pairs

Linearization/Delinearization, Z-Order, Hilbert and Morton Curves

Query the Live Tennis API from SQL - live matches, upcoming fixtures and player search as table functions

Call LLM APIs (OpenAI, Gemini, Cloudflare) directly from SQL with structured output support.

Query Grafana Loki from DuckDB in plain SQL, with label/time/line-filter pushdown into LogQL

Optimized-plan inspection and cross-system SQL transpilation

Largest-Triangle-Three-Buckets and related time-series downsampling aggregate functions.

Evaluate Lua scripts within queries

In-process LuaJIT UDFs for DuckDB — JIT-compiled, parallel (per-thread states), trusted sandbox, GROUP BY aggregates, LIST/STRUCT/MAP bridges, embedded Fennel compiler

libmagic/file utilities ported to DuckDB

Adds MARISA (Matching Algorithm with Recursively Implemented StorAge) trie functionality for DuckDB. MARISA is a static and space-efficient trie data structure that enables fast string lookups, prefix searches, and predictive text operations.

Read, analyze, and write Markdown files with block-level document representation and inline element support

Read MaxMind databases (GeoLite, GeoIP, and more)

Adds templating to DuckDB, enabling dynamic generation of text, HTML, and reports directly within SQL queries using the MiniJinja engine.

Interactive chart visualization for DuckDB - faster than pandas with Plotly-like charts

Lightweight, columnar-native ML for DuckDB — train + inference, 34 algorithms, zero Python deps

Connecting duckdb to the mlpack C++ machine learning library

Query AiM MP4, Cosworth PDS, MoTeC LD, and Racelogic VBOX telemetry directly in DuckDB

Read and write FIS Prophet files (.rpt, .prn, .fac)

Allows the consumption and production of the Apache Arrow interprocess communication (IPC) format, both from files and directly from stream buffers.

DuckDB extension for parsing, extracting, and analyzing domains, URIs, and paths with ease.

Read and write NSV (Newline-Separated Values) files

Validate, decode, and extract OAST (Out-of-Band Application Security Testing) domains in SQL

Provides IO observability to filesystem

A DuckDB extension for graph data analytics

Interact with LLMs with a simple DuckDB Extension

Access S3, local filesystem, and other storage services through Apache OpenDAL

Read and write Oracle Database with no Oracle client

A DuckDB extension for reading Apache ORC files, written in pure Rust. Supports all ORC primitives, struct, list, map types and all compression codecs (Zlib, Snappy, LZO, LZ4, ZSTD).

Read OpenStreetMap data from OSM XML and PBF files

Overture Maps helpers for DuckDB - geocoding, category normalization, spatial tiling, and reader macros for places, buildings, roads, and addresses

Automatic query privatization using the PAC Privacy framework

Query Apache Paimon tables directly from DuckDB

Exposes functions for parsing referenced tables and usage context from SQL queries using DuckDB's native parser.

DuckDB extension for querying Power BI Semantic Models with DAX.

Extension that allows parsing the data model embedded in PowerBI (pbix) files

Read and analyze PCAP network capture files directly in DuckDB using a native C++ implementation based on libpcap.

Extension that adds support for manipulating point cloud data using SQL.

Read, OCR, inspect, transform, and write PDFs in SQL

Graph algorithms for DuckDB — PageRank, shortest path, SCC, MST, and more via petgraph

Read PFC-JSONL compressed log files directly in DuckDB with block-level timestamp filtering

Picture-to-vector — image embeddings from SQL via tract-onnx (Rust). Bundles Meta's EUPE ViT-T/16; works with any ONNX vision model.

Lightweight geospatial indexing functions for DuckDB, including Geohash encoding, GEOMETRY decoding, bounding boxes, and neighbors.

Provides a spreadsheet-style pivot_table function

Read PLINK 2 genomics file formats and run common genetic analyses directly in SQL

SQL parsing and introspection for IDEs, editors, and developer tools

Support for PRQL, the Pipelined Relational Query Language

Support for PSQL, a piped SQL dialect for DuckDB

Pokemon data native in DuckDB

Embedded CPython UDFs for DuckDB — py_eval, py_agg, py_scan, py_register, py_call, py_map

Query Brendan Gregg folded CPU/allocation profiles as DuckDB tables

FastAPI-class HTTP framework inside DuckDB — CREATE ROUTE turns SQL into typed, validated endpoints; one process is DB + HTTP (+ PDF/renderer via companion extensions).

Bert-based embedding extension.

Query ISO 20022 (camt, pacs, pain) and SWIFT MT financial messages as SQL

DuckDB WireGuard Extension to Quack & Ducklake over Tailscale/Headscale

QuackStore - Smart Block-Based Caching for Remote Files. Speed up repeated queries on remote data with intelligent block-level caching.

This extension provides a cache for query conditions, which is useful for repeated-query workloads like metrics monitoring dashboards, log investigations, etc.

Reject queries before physical execution when estimated table scans exceed a configured row budget

DuckDB QuickJS Runtime Extension

Read (streaming) and write Qlik QVD files

Allow interaction with event buses like Websocket and Redis publish/subscribe servers.

Provides adds high-performance fuzzy string matching functions, powered by the RapidFuzz C++ library.

Raster analytics on Raquet files with QUADBIN spatial indexing, raster ingestion, and PostGIS-style functions

DuckDB extension for reading and writing geospatial raster data using SQL.

Perform rate and burst limit on filesystem operations

Schema-less JSON and OTEL ingestion shredder for DuckDB

Read and write RDF, query using SPARQL

Read DBF (dBase) files with support for general DBF formats and field types.

Read line-based text files with line numbers and efficient subset extraction. Supports glob patterns, line selection with context, and GitHub-style 'L12-24' fragments so a line range can be attached to any path, including URIs served by other extensions.

Redis compatible Client for DuckDB

A DuckDB extension that implements Predicate Transfer: derives filters from join keys and propagates them across the join graph of a multi-join query, so probe-side rows that can't survive downstream joins are pruned early.

ROCKET random convolutional kernel features for time-series classification

An Excel/WPS/OpenDocument Spreadsheets file reader for DuckDB

Read-only access to Salesforce orgs as DuckDB SQL tables over the official REST and Bulk APIs — OAuth refresh-token or JWT-bearer auth (credentials from inline options, environment variables, or an SFDX auth URL), native ATTACH, projection + predicate pushdown, COUNT pushdown, explicit server-side aggregates with GROUP BY, lazy/auto/Bulk transports with PK chunking (Bulk blob/base64 compatibility guard), a per-job API quota governor, parent + grandparent relationship STRUCT columns with diagnostics, queryAll (archived + deleted), Tooling-API fast schema, and metadata helpers (manual cache refresh, picklist values, record types).

A collection of virtual filesystems for working with scalars

Provides functionality for financial data-analysis, including data scanners for the Ethereum Blockchain and Yahoo Finance

Provides SE(3) rigid transformations (combinations of rotations and translations in 3D) as DuckDB scalar functions.

Fast XLSX file importer

Allow shell commands to be used for input and output

Query Slovenia's SiStat (SURS PX-Web API) open data directly from DuckDB with SQL.

Parse XML sitemaps from websites with automatic discovery via robots.txt

Snowflake data source extension - query Snowflake databases directly from DuckDB

Read logs from a Splunk search REST API into OTLP-shaped tables

Query SharePoint lists and Excel workbooks directly from DuckDB

Allows reading and writing files over SSH

DuckDB extension for reading data from SpatioTemporal Asset Catalogs (STAC) using SQL.

A statistical computing toolkit for DuckDB — descriptive statistics, hypothesis tests, R-style distribution functions (d/p/q/r), OLS regression with HC and cluster-robust standard errors, grammar-of-graphics charts (VISUALIZE → Vega-Lite), and first-class readers/writers for SAS, SPSS, and Stata files, all directly in SQL.

Add comprehensive statistical distribution functions to DuckDB, enabling advanced statistical analysis, probability calculations, and random sampling directly within SQL queries.

Read DuckDB database files written in a newer storage format than this DuckDB supports

Unified SQL access to Sudan's humanitarian, development, and geospatial data from 5 international APIs (World Bank, WHO, FAO, UNHCR, ILO)

Attach Superhuman Docs documents as DuckDB databases and query or modify their tables with SQL

Provides table functions to access system-level statistics for monitoring purpose

Enforces table-level access control via an allowlist mechanism, blocking queries to unauthorized tables with a PermissionException

Observability into DuckDB storage internals at the database, table, and column levels

Technical Analysis Library (TA-Lib) functions for DuckDB -- 100+ indicators for financial market analysis

Adds templating to DuckDB, enabling dynamic generation of text, HTML, and reports directly within SQL queries using the Tera engine.

Enables text-based data visualization directly in SQL queries, including ASCII/Unicode bar charts, density plots, and sparklines for lightweight analytics and dashboards.

3D solid processing for city models — enclosed volume, validation, distance, and measurement of polyhedral solids.

Efficiently standardizes scraped job titles to Bureau of Labor Statistics (BLS) titles using a high-performance TF-IDF algorithm.

Parse TOML documents into JSON

Native Trino-compatible scalar functions (trino_) for the cases where DuckDB's built-ins diverge from Trino on Unicode and byte-level inputs — ICU full case folding, code-point reverse, Java-whitespace trim, NFC normalize, and raw-bytes xxhash64/sha512/hmac_sha256

Query Apache TsFile table-model files directly from DuckDB

DuckDB Time-Sortable ID generator

Compressed vector search for DuckDB — Google TurboQuant quantization, IVF indexing, incremental CRUD, cosine/L2/dot

US address parsing and standardization (PAGC + Rust addrust)

DuckDB extension for routing and travel time calculations using Valhalla routing engine

VGI (Vector Gateway Interface) — DuckDB extension for extending DuckDB in any language over Apache Arrow IPC. No C++/compilation required.

Parse WARC (Web ARChive) records for Common Crawl data processing

Query Common Crawl and Wayback Machine web archive CDX APIs directly from SQL

Query Google Custom Search API directly from SQL with filter pushdowns

Comprehensive processing extension for web markup languages (XML and HTML) with SAX streaming for large files, intelligent schema inference, XPath-based data extraction, and HTML table parsing.

Speech-to-text transcription using whisper.cpp, OpenAI's Whisper model

Read and dissect PCAP files from DuckDB

Read YAML files into DuckDB with native YAML type support, comprehensive extraction functions, and seamless JSON interoperability

Measure-aware SQL implementing Julian Hyde's 'Measures in SQL' paper

Query Zarr array stores (v2 and v3) with SQL — read xarray-convention datasets as tables

Read Zeek network security monitor log files with automatic schema detection and type-aware parsing

Read files within zip archives