1bench — Modern vector database GUI client

One GUI for all
your vector databases

From dedicated vector engines like Qdrant and Weaviate to vector fields in Postgres (pgvector), Turso, and ClickHouse. Browse collections, run similarity search, and visualize embeddings without writing client-library code.

Apple Silicon — Also available for 7-day free trial · No credit card required

Your data stays yours

1bench runs entirely on your machine. Nothing is ever sent to us.

100% Local

Credentials and data never leave your machine.

Zero Telemetry

No usage tracking, no query logging, no analytics.

Direct Connection

Straight to your database, nothing in between.

Every vector-capable database, one app

Dedicated vector engines, all with the same interface.

QdrantQdrantTypesenseTypesenseMilvusMilvusWeaviateWeaviateChromaChroma

Plus vector fields in general-purpose databases

PostgreSQLPostgreSQLSupabaseSupabaseNeonNeonTursoTursoClickHouseClickHouseElasticsearchElasticsearchOpenSearchOpenSearchMeilisearchMeilisearch

Built for vector databases

1

Visual Similarity Search

Run similarity queries across every vector engine without writing Python or JS. Cosine, L2, or inner-product distance. Filter by metadata. Paste a raw vector or Find Similar by ID.

2

2D Vector Visualization

Project embeddings onto a 2D scatter plot. Color documents by any metadata field to reveal clusters. Click any point to inspect, right-click to Find Similar inline.

3

Find Similar

Right-click any document, vector, or search result to find its nearest neighbors. One-click semantic search from anywhere in the app, including the table, the scatter plot, and the collection browser.

4

Metadata Filtering

Build complex filters on vector metadata without touching client code. $eq, $ne, $gt, $lt, $in, $and, $or all work the same across every engine you connect to.

5

Cross-Engine Consistency

Learn one interface, use it everywhere. Qdrant collections, Milvus segments, Weaviate classes, and Chroma tenants all normalize into one mental model you already know.

6

Bulk Import & Export

Export collections as JSON with full documents, vectors, and metadata. Import from JSON to seed dev environments, backup production, or migrate between engines.

10x your vector database productivity

Stop switching between CLI and code. Manage everything visually.

Visual Query Builder

Search, filter, Find Similar, and raw-vector modes in a single UI. No client-library code required.

2D Vector Visualization

Scatter plot of embeddings with color-by-metadata. Click to inspect, right-click to Find Similar.

Find Similar

Right-click any document or vector to find its nearest neighbors. One-click semantic search everywhere.

Browse Collections

Paginate collections with full document, metadata, and vector visibility. Handles large datasets cleanly.

Metadata Filtering

Build complex metadata filters visually. $eq, $ne, $gt, $lt, $in, $and, $or supported across engines.

Inline Editing

Click any field to edit it in the table. Staged batch commit means no accidental writes.

Add Documents

Insert new documents with vectors and metadata through a clean visual form. No client code required.

Collection Management

Create, delete, and truncate collections safely with typed confirmation. Never lose data by accident.

Embedder Configuration

See each collection's embedding function, distance metric, and index config (HNSW, IVF, SPANN) inline.

Export & Import

JSON export with full documents, vectors, and metadata. Import to seed dev, backup, or migrate between engines.

Multiple Tabs

Work across vector engines and instances in parallel. Each tab keeps its own browsing and query state.

Cloud & Self-Hosted

Connect to Qdrant Cloud, Zilliz, Weaviate Cloud, Chroma Cloud, or self-hosted deployments with the same UI.

SSH Tunnel Support

Reach private-network vector instances securely through SSH tunnels. Load profiles from ~/.ssh/config.

Read-Only Mode

Enable read-only mode to safely explore production vector collections. Blocks all writes and deletes.

Query History

Every similarity query you run is automatically saved per connection. Search, re-run, or bookmark instantly.

Frequently asked questions

What is a vector database GUI?
A vector database GUI is a visual desktop client for browsing collections, running similarity search, editing documents and metadata, and inspecting embeddings without writing Python or JS client-library code. It replaces per-engine web consoles (Qdrant dashboard, Milvus Attu, Weaviate console) with one interface that speaks every vector database. 1bench is a modern vector database GUI for Mac, Windows, and Linux.
Do I need a different GUI for Qdrant, Milvus, Weaviate, and Chroma?
No. Each engine ships its own web console (Qdrant dashboard, Milvus Attu, Weaviate console, Chroma's hosted UI), which works fine if you only use one. If you run more than one vector engine (common in production RAG stacks), you end up learning four different tools and switching between four browser tabs. 1bench maps every engine's collections, segments, and classes into one consistent browser, editor, and similarity-search UI so the workflow stays identical across engines.
What is the best vector database GUI for Mac, Windows, or Linux?
A good vector database GUI should run natively on your OS, connect to both cloud and self-hosted instances, run similarity search with metadata filters, and visualize embeddings on a 2D scatter plot. Per-engine dashboards each miss at least one of these: Qdrant's is web-only, Milvus Attu is JVM-based, Chroma's dashboard is Cloud-only. 1bench delivers all four in a single native download for Mac, Windows, and Linux, and works across every major vector engine.
Can I run a vector database GUI locally or self-hosted?
Yes. Every dedicated vector engine (Qdrant, Milvus, Weaviate, Chroma) ships a Docker image or standalone binary that runs on your laptop, and 1bench connects to them over the same connection form as their cloud counterparts. It also supports SSH tunnels for reaching private-network instances, plus a read-only mode so you can safely explore production data without risk of accidental writes.
Does a vector database GUI work with pgvector and vector-in-general-purpose databases?
Yes, if the GUI treats them as first-class. 1bench connects to vector fields in Postgres (pgvector), Turso, ClickHouse, Elasticsearch, OpenSearch, and Meilisearch, and exposes the same similarity search, Find Similar, and metadata filtering it does for dedicated vector engines. That way you can start with pgvector inside your existing Postgres and migrate to Qdrant or Milvus later without changing tools.

Ready to manage your vector databases visually?

1bench connects to every vector database we support, both cloud and self-hosted. Start your 7-day free trial today.

View Pricing