01
Overview
Is everything okay?
Clarity across
databases.
One workspace for your databases and vectors. Safe Mode for production.
Current release is freeVersion 1.0.0 · macOS, Windows and Linux
Desktop app · v1.0.0Recern WorkspaceA desktop workspace for PostgreSQL, MySQL, Microsoft SQL Server, MongoDB, ClickHouse, Cassandra, Qdrant, OpenSearch, SQLite, Redis, Valkey, Dragonfly and Recern Vector: activity, queries, indexes and findings, with Safe Mode for production.Explore the workspace Open source · v0.4.0Recern VectorRecern Vector 0.4.0 combines embedded vector storage with a single-node HTTP server and portable Python, Rust and Node.js clients for Recern and Qdrant.Explore Recern Vector
01 / A clearer perspective
Understand what’s happening,
not just what’s stored.
Traditional database clients are built around tables, rows and query editors. Recern brings data, activity and performance into one workspace.
02
Data
What is stored here?
03
Activity
What is happening right now?
04
Performance
What is slow or expensive?
05
Console
Talk directly to the database.
Other database tools help you manage databases. Recern helps you understand them.
02 / Built around your database
Different databases.
Native understanding.
Each engine has its own way of working. Recern respects that — with workflows built around its runtime, not a generic interface.
AvailableCassandraQdrantOpenSearchComing laterAzure SQLDynamoDBDatabricks SQLOracle DatabaseIBM Db2SAP HANAAll databases and what comes next
Insights
See what deserves
attention.
Recern reads each server's own statistics and turns them into findings: a blocked session, an unused index, a collection full of deleted vectors. Each one leads to the view that explains it.
Inspect. Understand. Decide.
03 / Production-aware by design
Built for production.
Production connections start in Safe Mode. PostgreSQL and MySQL use read-only transactions; SQLite opens read-only and blocks unsafe operations; Redis refuses commands that write. TLS and SSH tunnels are available for server connections. MongoDB, ClickHouse, Cassandra, Qdrant, OpenSearch and Microsoft SQL Server consoles are read-only in this version. TLS support through SSH depends on the engine; see the connection documentation. Use a dedicated read-only account and query limits; Safe Mode does not replace database permissions or prevent expensive reads.
Safe Mode for productionRead-only consoleSSH tunnels with host key checksTLS and Keychain passwords
PRODUCTIONSafe Mode ON
Confidence before every command.
04 / Recern Vector Open source · v0.4.0
A vector database
in a single file.
Recern Vector 0.4.0 combines embedded vector storage with a single-node HTTP server and portable Python, Rust and Node.js clients for Recern and Qdrant.
- Portable Python, Rust and Node APIs
- Recern HTTP server
- Verified, resumable migration
- Stable files, WAL and int8
MIT OR Apache-2.0 · v0.4.0
import recern_vector as rv
db = rv.Database.open_or_create("docs.rvec")
docs = db.create_collection("docs", dim=384, metric="cosine")
docs.upsert_many(ids, embeddings, metadatas)
hits = docs.search(query, k=5, filter={"lang": "en"})
docs.explain(query, k=5) # strategy, visited nodes, timeRecern Vector 0.4.0 · 2026-11-02 · embedded f32 and int8; HTTP and Qdrant are outside this measurement.
7,712/s
queries/s on SIFT1M at recall ≥ 0.95
618/s
queries/s on GloVe-100 at recall ≥ 0.95
102.7s
to build the SIFT1M f32 index · 4 threads
Open the file in Recern Workspace
Workspace inspects pgvector columns and indexes, Redis search indexes and vector sets when the server supports them, and Recern Vector snapshots with committed WAL changes. Recall checks are available for Recern Vector collections, including f32 and int8.
A useful first session
Follow a short guide from a real question to an answer in Recern.
Why is this query waiting?
Follow a PostgreSQL lock chain to the transaction that holds it up.
Does this vector query match its index?
Check pgvector columns, distance operators, settings and index usage.
Build your first vector search
Create a local collection, filter results and inspect how the search ran.
05 / Questions / Answers
Recern,
explained clearly.
Direct answers about the workspace, supported databases and Recern Vector.
01What is Recern?
Recern is a desktop workspace for databases. It brings activity, performance and the findings that need attention into one clear view, and keeps production connections safe by default. Alongside it we build Recern Vector, an open-source embedded vector database.
02Which databases does Recern support?
Recern Workspace 1.0.0 supports PostgreSQL with pgvector, MySQL, Microsoft SQL Server, MongoDB, ClickHouse, Cassandra, Qdrant, OpenSearch, SQLite, Redis, Valkey, Dragonfly and Recern Vector files. Microsoft SQL Server is available with SQL authentication and verified TLS.
03What about vector search?
Workspace inspects pgvector columns and indexes, Redis search indexes and vector sets when the server supports them, and Recern Vector snapshots with committed WAL changes. Recall checks are available for Recern Vector collections, including f32 and int8.
04How does Recern approach production databases?
Production connections start in Safe Mode. PostgreSQL and MySQL use read-only transactions; SQLite opens read-only and blocks unsafe operations; Redis refuses commands that write. TLS and SSH tunnels are available for server connections. MongoDB, ClickHouse, Cassandra, Qdrant, OpenSearch and Microsoft SQL Server consoles are read-only in this version. TLS support through SSH depends on the engine; see the connection documentation. Use a dedicated read-only account and query limits; Safe Mode does not replace database permissions or prevent expensive reads.
05Which platforms does Recern run on?
Recern is a desktop app for macOS, Windows and Linux. Builds appear on the download page as soon as they are released.