HeadsUpAI

DigitalOcean launches unified data layer to simplify AI agent retrieval

DigitalOcean launched the Data & Learning layer to ground AI applications in real-time data. The release moves Knowledge Bases to general availability, providing a managed RAG (retrieval-augmented generation—grounding AI with external knowledge) pipeline that automates ingestion. It also introduces managed Weaviate in private preview for vector search.
Knowledge Bases Pricing
$19.60 per month
Embedding Cost
$0.02 per 1M tokens
SQL Scaling Speed
Minutes (vs. hours)
Vector Engine
Weaviate (Private Preview)
SQL Editions
PostgreSQL and MySQL Advanced

This update addresses the "auth glue" problem where developers stitch together separate vendors for transactional records and semantic search. By co-locating storage with its inference platform, DigitalOcean eliminates egress fees between data and models. This follows a trend of cloud providers like Vercel integrating managed search for agents.

Users can now deploy PostgreSQL and MySQL Advanced Edition in public preview, featuring proxy-based failover and capacity scaling that takes minutes. Knowledge Bases starts at $19.60 per month with embedding tokens at $0.02 per million. The system supports the Model Context Protocol, allowing agents to connect to data sources without custom code.

View the full update on digitalocean.com
DigitalOcean
DigitalOcean
@digitalocean
X

Building AI agents? You need retrieval, not just storage. 🤖 DigitalOcean just launched Knowledge Bases (GA), Managed Weaviate (Private Preview), & PostgreSQL & MySQL Advanced (Public Preview)—a full data layer for the Inference Era. Managed RAG. Open-source vectors. Databases that scale in minutes instead of hours. ✨

4retweets8likes
View on X

Still wondering? A few quick answers below.

DigitalOcean Knowledge Bases is a fully managed RAG pipeline that handles document ingestion, chunking, and embedding. It allows developers to upload documents and select an embedding model to create a citation-backed knowledge source for AI agents without managing the underlying vector infrastructure or retrieval logic.

The Advanced Edition for PostgreSQL and MySQL is designed for high-growth AI workloads requiring greater resilience. It features proxy-based failover to keep applications alive during node failures and allows for capacity scaling in minutes rather than hours, providing a more stable foundation for agents accessing transaction logs.

Managed Weaviate is currently in Private Preview. It offers the standard open-source vector engine as a managed service within the DigitalOcean ecosystem. This allows teams to maintain control over their schemas and chunking strategies while benefiting from unified billing and zero egress fees between storage and inference.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Share this update