infinity

The AI-native database built for LLM applications, providing incredibly fast hybrid search of dense vector, sparse vector, tensor (multi-vector), and full-text.

What is infinity?

The AI-native database built for LLM applications, providing incredibly fast hybrid search of dense vector, sparse vector, tensor (multi-vector), and full-text.

Overview of infinity

What it is

Infinity is an AI-native database designed for LLM applications. It provides hybrid search capabilities across dense vector, sparse vector, tensor (multi-vector), and full-text data types. The database is built to support retrieval-augmented generation (RAG) applications such as search, recommenders, question-answering, conversational AI, copilot, and content generation.

Key features

  • Incredibly fast: achieves 0.1 milliseconds query latency and 15K+ QPS on million-scale vector datasets; 1 millisecond latency and 12K+ QPS in full-text search on 33M documents.
  • Powerful search: supports hybrid search of dense embedding, sparse embedding, tensor, and full text, with filtering; includes rerankers such as RRF, weighted sum, and ColBERT.
  • Rich data types: supports strings, numerics, vectors, and more.
  • Ease-of-use: intuitive Python API, single-binary architecture with no dependencies, and can be embedded in Python as a module.

Use cases

Infinity is suitable for LLM applications including search, recommenders, question-answering, conversational AI, copilot, content generation, and various RAG (retrieval-augmented generation) applications.

Getting started

Prerequisites: CPU x86_64 with AVX2 support; Linux (glibc 2.17+), Windows 10+ with WSL/WSL2, or macOS; Python 3.11+. Install Infinity server via Docker: bash sudo mkdir -p /var/infinity && sudo chown -R $USER /var/infinity docker pull infiniflow/infinity:nightly docker run -d --name infinity -v /var/infinity/:/var/infinity --ulimit nofile=500000:500000 --network=host infiniflow/infinity:nightly Install Infinity client: pip install infinity-sdk==0.7.4 Example vector search (Python): python import infinity infinity_obj = infinity.connect(infinity.NetworkAddress('', 23817)) db_object = infinity_object.get_database('default_db') table_object = db_object.create_table('my_table', {'num': {'type': 'integer'}, 'body': {'type': 'varchar'}, 'vec': {'type': 'vector, 4, float'}}) table_object.insert([{'num': 1, 'body': 'unnecessary and harmful', 'vec': [1.0, 1.2, 0.8, 0.9]}]) table_object.insert([{'num': 2, 'body': 'Office for Harmful Blooms', 'vec': [4.0, 4.2, 4.3, 4.5]}]) res = table_object.output(['*']).match_dense('vec', [3.0, 2.8, 2.7, 3.1], 'float', 'ip', 2).to_pl() print(res) For binary deployment or building from source, see the repository README.

License and pricing

The project is licensed under Apache-2.0 and is open source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/infiniflowinfinity.svg)](https://yunodb.com/tools/infiniflowinfinity?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/infiniflowinfinity?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/infiniflowinfinity.svg" alt="Featured on YunoDB" height="20"></a>