Weaviate 1.38 Release

2026-06-25 · Source: Weaviate Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Intermediate, medium

Summary

Weaviate v1.38, released on June 25, 2026, is now available open-source and on Weaviate Cloud, featuring significant enhancements. The HFresh disk-based vector index and the built-in MCP Server have reached general availability. HFresh, inspired by SPFresh, uses on-disk postings and an in-memory HNSW index, providing predictable latency and low memory usage for streaming workloads, and includes RQ-1 quantization. The MCP Server, a Streamable HTTP endpoint at "/v1/mcp", allows LLMs and AI agents to interact directly with Weaviate for schema inspection, hybrid searches, and object upserts, with runtime-configurable write access. Async replication has been re-architected for cluster-wide operation from a single scheduler, running by default on all replicated collections. New preview features include the Boost API for query-time rescoring (gRPC-only, up to 20 conditions) and Nested Object Filtering, enabling filtering on nested object properties via dotted paths. The release also brings production-ready replica movement, a `text2vec-digitalocean` module, and various performance improvements.

Key takeaway

For MLOps Engineers managing large-scale vector databases, Weaviate v1.38 offers critical advancements. You should evaluate HFresh for memory-efficient handling of streaming data, especially for collections growing into billions of vectors. Consider enabling the MCP Server to streamline direct integration with LLMs and AI agents, enhancing automation. Utilize the new cluster-wide async replication for improved data consistency and simplified operations across your replicated collections. Explore the Boost API for fine-grained query-time result re-ranking.

Key insights

Weaviate v1.38 enhances vector search with disk-based indexing, direct AI agent interaction, and improved replication for scalable, flexible operations.

Principles

• Disk-based indexing reduces memory for large datasets.
• Direct API access simplifies AI agent integration.
• Centralized async replication improves cluster consistency.

In practice

• Configure HFresh for streaming workloads with billions of vectors.
• Enable MCP Server for LLM/AI agent direct schema and query access.
• Use Boost API to re-rank search results without dropping any.

Topics

• HFresh Vector Index
• MCP Server
• Async Replication
• Boost API
• Nested Object Filtering
• LLM Integration
• Vector Quantization

Code references

• weaviate/weaviate

Best for: AI Architect, AI Engineer, Machine Learning Engineer, MLOps Engineer

Related on AIssential

• Weaviate 1.37 Release — Weaviate "v1.37" transforms the vector database into an active AI agent memory and enhances core search, backup, and multimodal capabilities.
• Vector Data Types & Native Indexing: The ANN Revolution We Didn’t Know We Needed — Native vector types and indexing in SQL engines enable efficient ANN searches, fundamentally changing data management for AI workloads.
• CoreWeave AIRA: The autoresearch loop for continuous improvement — Arya automates ML research by analyzing past runs, proposing new trials, and launching them for continuous improvement.
• Vector Search at Scale: The Production Engineer's Guide — Efficient vector search at scale requires sub-linear time, memory compression, and approximate results.
• How Vector Databases Use Specialized Indexes (HNSW, IVF, PQ, and Annoy) — Vector databases use specialized Approximate Nearest Neighbor (ANN) indexes to achieve millisecond similarity search across millions of high-dimensional vectors.
• Build a Coding Assistant with Weaviate MCP: RAG over Code & Docs — Weaviate's built-in MCP server simplifies RAG for coding assistants, enabling hybrid search over code and docs.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Weaviate Blog.