Menu

Post image 1
Post image 2
Post image 3
1 / 3
0

Kreuzberg & SurrealDB: from unstructured documents to hybrid retrieval

DEV Community·Mark Gyles·21 days ago
#Jd7s5ecQ
Reading 0:00
15s threshold

Author: Ignacio Paz We’re excited to share a new partner integration: kreuzberg-surrealdb , a connector that bridges the Kreuzberg document intelligence framework directly into SurrealDB. This integration was created by the Kreuzberg team and we are excited to have this functionality available now in SurrealDB. Kreuzberg extracts, chunks, and generates embeddings from 88+ document formats, while SurrealDB provides a multi-model database for AI applications, combining documents, graphs, vectors, and full-text search in a single system. Together, they make it easy to build document search and RAG pipelines. What the integration does kreuzberg-surrealdb handles the full ingestion workflow: Automatic schema setup Content deduplication using SHA-256 hashing Storage and indexing in SurrealDB Documents ready for search immediately after ingest The integration supports two modes: DocumentConnector:indexes full documents for BM25 keyword search.…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.

Read More