Deploy AI agents with persistent memory.Ship them anywhere.

Sub-15ms retrieval on your own hardware. On-premises, at the edge, air-gapped, or cloud. Same architecture throughout.

Start building free Read the docs
pip install actian-vectorai
13ms · 5 results · 99.1% recall
8,401 downloads · Trusted by
IntuitLufthansaSabreAmerican AirlinesConocoPhillipsAcademy Bank IntuitLufthansaSabreAmerican AirlinesConocoPhillipsAcademy Bank

What developers are building
with VectorAI DB

V
Trace
Semantic search for missing persons cases. Bridges the vocabulary gap between families and forensic records, fully on-device.
View project
V
Revvec
Air-gapped RAG pipeline for defense, aerospace, and pharma. Runs on a single Mac, no internet required.
View project
V
FIX-FIRST
Offline maintenance copilot for industrial technicians. Search manuals and incident history by error code, photo, or voice note.
View project
V
Community plugin
LangChain + VectorAI DB
Agent memory that persists across sessions without cloud dependencies.
Read the guide
V
Community plugin
LangGraph + VectorAI DB
Agent state that survives disconnected environments.
Read the guide
V
Community plugin
CrewAI + VectorAI DB
Isolated, persistent memory for every agent in a multi-agent crew.
Read the guide
V
Community plugin
LlamaIndex + VectorAI DB
Local RAG pipeline with no cloud vector service. Runs on-premises or at the edge.
Read the guide
See all community builds →

Most vector databases were built for
human query patterns. AI agents are different.

Cloud latency breaks agent reasoning loops
Agents hit memory on every step. Cloud round-trips add 200 to 400ms per call, breaking the latency budget before the model reasons.
+340ms
+280ms
+390ms
six-step reasoning chain · 500ms budget
Regulated deployments require architectural compliance
In regulated environments, a vendor contract is not enough. Agent memory has to stay on infrastructure you control.
your infrastructure
Cloud-only architecture leaves entire deployments unbuilt
Factory floors, hospitals, and field devices need data on local hardware. Cloud-first databases have no answer.
Deployment status● live
Cloud vector DB✗ No connection
VectorAI DB✓ Running · offline
Agent queries✓ 13ms p99
47,382queries served

Numbers that hold up under load

Throughput
0
745.2 QPS on 10M vectors at 768 dimensions. More than 22x faster than Milvus and Qdrant on identical hardware.
Recall accuracy
0%
Retrieval accuracy holds as your agent's memory grows, at production volume, weeks into a live deployment.
11 of 12 true neighbours returned
Latency p99
0ms
13ms p99 at 99% recall, measured on self-hosted hardware, where your agents actually run.

What you get

01
Memory that survives every session
Recall across sessions, scaled with usage, inside your perimeter.
02
Monitor it from the tools you already run
Native Prometheus metrics in your existing Grafana dashboards.
03
Parallel queries without the slowdown
Parallel reasoning steps and memory lookups, recall intact.
04
Ship the same build everywhere
Laptop to data center with zero rewrites, any hardware.
05
Pass review without re-architecting
GDPR, HIPAA, and residency hold by where the data sits.

Watch VectorAI DB run

Watch a full agent deployment from zero to first query.

2:01
PRODUCTION DEMO · 2:01

Get started with
your framework

Drop it into the stack you already work in.

Start building free
agent_memory.py
View all integrations →

Run it however you ship

Python and JavaScript SDKs, a REST API, and a single Docker container that runs the same on a laptop, an edge box, or a Kubernetes cluster.

Python & JavaScript SDKs
Idiomatic clients for both languages with the same API surface. Bring your own embedding model.
# Python
pip install actian-vectorai

// JavaScript
npm install @actian/vectorai
Docker in one command
Clone the repo and bring the database up with Docker Compose. Running locally in under a minute.
# Start the database
docker compose up

# Database ready on :8080
Kubernetes, Helm & Terraform
The same container is Kubernetes, Helm, and Terraform compatible. Deploy to a cluster with the tooling you already use.
KubernetesHelmTerraformDocker Compose
REST API
Every operation is available over a REST API, so any language or runtime can talk to VectorAI DB directly.
# Any language, any runtime
curl localhost:8080/v1/collections

How VectorAI DB compares

An honest look at where each vector database runs and how it performs. Throughput measured on 10M vectors at 768 dimensions, identical hardware.

 VectorAI DBPineconeQdrantWeaviate
Throughput at 10M vectors745 QPS~33 QPS~34 QPSself-hosted varies
Runs on resource-constrained edge hardware (Jetson, Pi)YesNoLimitedLimited
Same API from prototype to productionYesCloud onlyYesYes
On-premises and air-gapped deploymentYesNoYesYes
Fully managed cloud optionSelf-managedYesCloudCloud

Pinecone and Qdrant throughput figures from independent 2026 benchmarks at comparable scale. VectorAI DB is built for teams that run where their data lives. If you want a hands-off managed cloud, the others are a strong fit.

Who builds on VectorAI DB

Teams shipping where the data lives and the network can not always reach.

{edge}
Edge AI engineers
Inference and memory on the same hardware. Runs natively on NVIDIA Jetson, Raspberry Pi, and industrial edge servers.
→ Jetson · RPi · Edge servers
{air-gapped}
Manufacturing teams
Serves queries during network outages and syncs to the data center when connectivity returns.
→ Air-gapped · Plant floors · Production lines
{hipaa}
Healthcare organizations
Patient data stays inside your perimeter throughout every query.
→ Hospital DCs · Clinic servers · Research
{hybrid}
Platform engineers
Consistent across hybrid, edge, and multi-site infrastructure regardless of connectivity, compliance, or hardware constraints.
→ Hybrid · Edge + cloud · Multi-site

Common questions

What is VectorAI DB?

Actian VectorAI DB is a vector database built for AI agents, giving them persistent memory and low-latency retrieval on infrastructure you own.

How is VectorAI DB different from Pinecone, Qdrant, and Weaviate?

Most vector databases were designed for cloud-native deployments and human-scale query patterns. VectorAI DB is built for the concurrent, high-frequency behavior of agent fleets and runs consistently across edge, on-premises, air-gapped, and cloud environments without architecture changes. Pinecone is cloud-only. Qdrant and Weaviate support self-hosted deployment but are designed primarily for cloud-scale workloads, not edge and disconnected environments.

What agent frameworks does VectorAI DB support?

VectorAI DB integrates natively with LangChain, LlamaIndex, and Hugging Face. Python and JavaScript SDKs are available. Community plugins extend support to CrewAI, LangGraph, and Mastra. Full integration docs are at docs.vectoraidb.actian.com.

How does VectorAI DB handle concurrent queries from a fleet of agents?

VectorAI DB delivers 745.2 QPS on 10M vectors at 768 dimensions, more than 22x faster than Milvus and Qdrant on identical hardware. It is built for the parallel, high-frequency query behavior that agent fleets generate, not for the sequential patterns of human-facing applications.

What does the free tier include?

The free tier is the VectorAI DB Community Edition. Free forever, no billing setup. It includes 5K vector capacity, deploys on local development machines, and comes with community support.

What indexing algorithms does VectorAI DB support?

VectorAI DB supports HNSW and other ANN indexing methods for low-latency, high-accuracy search at scale.

What embedding models does VectorAI DB support?

VectorAI DB is model-agnostic and works with embeddings from any provider: OpenAI, Anthropic, Cohere, Hugging Face, or any custom model you bring.

Does VectorAI DB support multi-modal embeddings?

Yes. VectorAI DB stores vector embeddings from text, images, audio, and video.

Deploy AI agents with persistent memory.
Ship them anywhere.

No credit card required. From install to your first query in under five minutes.

Start building free Get a demo

Start free, no billing setup

Everything you need to prototype agent memory on your own machine, today.

Billing-free access
Download and start using VectorAI DB immediately. No billing setup required.
Vector indexing and search
ANN indexing (HNSW), vector search APIs, and monitoring to support prototypes up to 5,000 vector embeddings.
Fast local setup
Install, index data, and run queries in minutes with clear docs, quickstarts, and code samples.
Developer community
Join our Discord to ask questions, share projects, and get help from other builders.
Join the Discord

From the engineering blog

Deep technical writing on agent memory, retrieval, and running vector search off the cloud.

BENCHMARKS
How to Evaluate Vector Databases in 2026
Why production selection demands real-world testing beyond vendor benchmarks: tail latency, concurrent filtering, and total cost of ownership.
Read the post
RAG
Turn Your Internal Docs into an AI-Powered Knowledge Base
A RAG walkthrough with vector search: chunk and index your docs, retrieve the most relevant chunks, then generate grounded answers.
Read the post
ON-PREM
3 Lessons From Building With VectorAI DB at Hacklytics 2026
How three healthcare teams used vector search to solve vocabulary mismatch, enable explainable retrieval, and deploy compliant local stacks.
Read the post