The AI pipeline you're looking for is actually a database.
Search. RAG. Knowledge graph. AI model garden.
All in the database.
See it in action
A simple REST API for hybrid search and RAG pipelines
1# Create a table with vector embeddings
2curl -X POST http://localhost:8080/tables \
3 -H "Content-Type: application/json" \
4 -d '{
5 "name": "docs",
6 "indexes": [{
7 "type": "aknn_v0",
8 "fields": ["content"],
9 "embedder": {
10 "provider": "termite",
11 "model": "bge-small-en-v1.5"
12 }
13 }]
14 }'Why AntflyDB
Replace the patchwork
Building AI search means stitching together databases, queues, APIs, and custom code. AntflyDB replaces the entire stack.
Each service requires separate hosting, monitoring, billing, and error handling.
Everything you need
A complete platform for building AI-powered search and retrieval applications
Hybrid Search
BM25 + vector similarity with Reciprocal Rank Fusion. Get the best of both lexical and semantic search.
Multi-Raft Consensus
Separate consensus groups for metadata and storage. Scale your cluster without bottlenecks.
Multimodal
Index and search text, images, audio, and video. CLIP embeddings for cross-modal retrieval.
Local ML Inference
Termite runs embeddings locally with ONNX models. Your data never leaves your infrastructure.
Kubernetes Native
Official operator with autoscaling, rolling updates, and automated lifecycle management.
Hardware Accelerated
SIMD optimizations for AVX-512, NEON, and SME. Blazing fast on any architecture.
Get started in minutes
From zero to hybrid search in three commands
Install
brew install antflydb/antfly/antflyPull Models
antfly termite pull bge-small-en-v1.5Run
antfly swarmLike Ollama, but for all AI models. Run embeddings, chunking, and reranking locally with ONNX-optimized models.