<50ms query latency

+200% relevant context

The only vector database with a feedback loop

Connect your data. Train your models. Query text, user or session context and retrieve relevant results in milliseconds.

Get $100 credit

Book a demo

$100 free credits. No credit card required.

ShapedQL

One query. Any use case.

Retrieve by text, user ID, or item ID — whatever context you have.

Retrieve

Hybrid search across multiple indexes

Filter

Hard constraints & business rules

Score

ML models & value functions

Reorder

Diversity & exploration

Agent context

01
02
03
04

                        SELECT
                         doc_id, title, content
                      
FROM

                          semantic_search("How do I authenticate the API?"),
                      

                          keyword_search("How do I authenticate the API?")
                      

05

                        WHERE verified = true

06
07
08
ORDER BY

                          colbert_v2(item, "How do I authenticate the API?") +
                      

                          recency_score(item)
                      

09
10
REORDER BY

                          diversity(strength=0.3)
                      

Unlike a vector store, Shaped knows who's asking. Results are personalised per user and get better with every interaction.

Try this query

Personalized hybrid search

01
02
03
04

                        SELECT
                         title, description
                      
FROM

                          semantic_search("wireless headphones"),
                      

                          keyword_search("wireless headphones")
                      

05
06
07
ORDER BY

                          colbert_v2(item, "wireless headphones") +
                      

                          click_through_rate_model(user, item)
                      

Blend semantic and keyword search, reranked by what this user actually clicks.

Try this query

Also available via Python SDK, TypeScript SDK, or MCP. See docs

Built for decisions, not documents

Legacy RAG stacks require endless duct tape. Shaped unifies retrieval, ranking, and learning in one query.

Traditional agent stacks

$1.50 per answer

Retries from bad context add up fast.

$0.03 per answer

50x cheaper. No retries. No surprises.

Retry loops

Hallucinate, retry, repeat.

No retry loops

Right results first time, every time.

Sliced data

5 systems. 5 calls. Bad context.

Unified context

One call. Every data source. Instant.

Static embeddings

Never learn. Never improve.

Feedback loop

Every interaction makes results better.

5 services

Pinecone + Cohere + Redis + pipelines + glue.

One query

Replace your entire retrieval stack.

Text only

Can't retrieve by user or item.

Any context

Text, user ID or item ID.

Similarity matching

No business logic. No personalisation.

Ranks for decisions

Behavioural signals + business rules.

Build everything yourself

Pipelines, connectors, maintenance.

30+ native connectors

Plug in. Stream. Done.

Replace your agent stack

Three-layer architecture

A unified, queryable relevance engine

Shaped is an end-to-end relevance engine designed for real-time personalization and agent memory.

Query layer

Real-time retrieval & ranking

ShapedQL
SQL
interface

Ranking
Multi-stage
pipeline

Results
<50ms latency

Learn more

Intelligence layer

ML models & embeddings

Embeddings
Dynamic
vectors

Training
Continuous
learning

Models
User + item
context

Learn more

Data layer

30+ connectors

Connectors
Batch +
streaming

Tables
Unified
schemas

Views
SQL
transforms

Learn more

Proven results

Trusted by leading teams

Product and engineering teams use Shaped to drive engagement and revenue

Read case studies

The only vector database with a feedback loop

One query. Any use case.

Agent context

Personalized hybrid search

Built for decisions, not documents

A unified, queryable relevance engine

Query layer

Intelligence layer

Data layer

Trusted by leading teams

+79% D6 Retention

+4.9% Bookings

+9% Purchases

+10% Watch Time

+8% CTR

+13% Engagement

+10% Email CTR

+382% Diversity

+16% AOV

2.2X Conversion

+131% Watch Time

10x

Increase in experimentation velocity

~7 days

Time to first experiment

Connect your stack. Unify your data.

Built for every use case

For you feeds

Search & discovery

Agent retrieval

Similar items

Personalized email

AI assistants

30+ Native Connectors

Enterprise-grade security

Enterprise ready

Get started this week