LanceDB Studio for VS Code
English | 简体中文
English
LanceDB Studio is a VS Code extension for browsing, filtering, editing, exporting, and searching local or S3-compatible LanceDB databases.
Community Project Disclaimer: LanceDB Studio is a community extension and is not affiliated with, sponsored by, or endorsed by LanceDB. The LanceDB name and related trademarks belong to their respective owners.
UI Preview
Add Local Connection
Version & History
Browse Tables
SQL Condition Autocomplete
Features
- Save, reopen, and rename local, Amazon S3, and MinIO connections
- Automatically detect the parent database directory when selecting a single
.lance table folder
- Browse data tables and Arrow Schema fields in the Activity Bar
- Filter data using raw SQL conditions and structured field filters (combined with
AND)
- SQL autocomplete for fields, functions, and operators based on the current table schema
- Sort by fields without loading the entire table into memory
- Schema-driven dynamic forms to insert, view, edit, and delete single rows
- Copy cell contents or full row JSON
- Perform vector search with query vectors, metadata filters, Top K, and distance metrics
- Display the
_distance score used to rank vector-search results
- Find records similar to the currently selected row
- View, create, replace, and drop vector, scalar, and full-text indexes with Schema compatibility checks
- Export complete filtered result sets as CSV, JSONL, Parquet, or Arrow IPC
- Combine full-text and vector retrieval with Hybrid Search and reciprocal-rank or weighted fusion
- Use the built-in OpenAI-compatible embedding provider or register custom embedding providers and rerankers through the extension API
- Save named queries, reopen recent filter history, inspect table versions, and restore an earlier version safely
- Inspect table size, row and fragment statistics, index coverage, and schema-aware index suggestions
Quick Start
- Open LanceDB Studio in the Activity Bar.
- Select Add Local Connection and choose a directory, or select Add S3/MinIO Connection and enter an
s3:// URI plus optional endpoint and credentials.
- Expand the connection and open a table.
- Browse and maintain data using SQL conditions, field filters, pagination, and row operations.
Security & Privacy
- The extension does not collect telemetry. S3 connections access only the object-storage location you configure.
- Hybrid Search sends query text to an embedding provider only when you explicitly select that provider; the manual-vector provider is fully local.
- The Webview never imports or accesses LanceDB directly; all database operations run inside the Extension Host.
- Single-page data limit is capped at 500 rows.
- Structured filter and sort fields are validated against the table Schema before execution.
- Row writes use LanceDB's internal row id together with the table version loaded by the UI; stale writes are rejected.
- Mutations are serialized per table and verify the actual affected-row count after execution.
- Edit forms submit changed fields only, and 64-bit integers cross the Webview boundary as exact decimal text.
- The most recent row mutation can be undone while no newer table version exists.
- Explicit confirmation is required before deleting tables, rows, or indexes.
Embedding configuration
The built-in OpenAI-compatible provider uses lancedbStudio.embedding.endpoint, lancedbStudio.embedding.model, and the API-key environment variable named by lancedbStudio.embedding.apiKeyEnvironmentVariable (default: OPENAI_API_KEY). Other VS Code extensions can register embedding providers and rerankers through the API returned by this extension's activate function.
Current Scope
Local filesystems and S3-compatible object stores are supported. Managed LanceDB Cloud accounts and data import are not included yet.
Author & License
简体中文
LanceDB Studio 是一个本地优先的 VS Code 插件，用于浏览、筛选、编辑和搜索本地 LanceDB 数据库。
社区项目声明： LanceDB Studio 是社区插件，与 LanceDB 官方无隶属、赞助或背书关系。LanceDB 名称及相关商标归其各自权利人所有。
界面预览
添加本地连接
版本 & 历史操作
浏览表格
SQL 条件自动补全
功能
- 保存并重新打开本地 LanceDB 连接
- 选择到单个
.lance 表目录时，自动识别其上一级数据库目录
- 在 Activity Bar 中浏览数据表和 Arrow Schema 字段
- 使用原始 SQL 条件和结构化字段条件筛选数据，两者以
AND 组合
- 根据当前表结构补全 SQL 字段、函数和运算符
- 无需加载整张表即可进行字段排序
- 根据 Schema 自动生成表单，新增、查看、修改和删除单行数据
- 复制单元格内容或整行 JSON
- 使用输入向量、元数据条件、Top K 和距离算法执行向量检索
- 显示向量检索结果的排序依据
_distance
- 查找与当前选中数据行相似的记录
- 查看、创建、替换和删除向量索引、标量索引及全文索引，并按 Schema 校验兼容性
快速开始
- 在 Activity Bar 中打开 LanceDB Studio。
- 选择 Add Local Connection，然后选择包含一个或多个
.lance 表目录的 LanceDB 数据库目录。
- 展开连接并打开数据表。
- 使用 SQL 条件、字段筛选、分页和行操作浏览或维护数据。
安全与隐私
- 插件不收集遥测数据，也不会主动上传数据库内容。
- Webview 不会导入或直接打开 LanceDB，所有数据库操作均在 Extension Host 中执行。
- 单页数据量上限为 500 行。
- 结构化筛选字段和排序字段会先与数据表 Schema 校验。
- 行写入同时使用 LanceDB 内部行标识和界面加载时的数据表版本；陈旧写入会被拒绝。
- 同一数据表的写入会串行执行，并在执行后核对实际影响行数。
- 编辑只提交实际变化的字段，64 位整数以精确十进制文本跨越 Webview 边界。
- 在没有更新数据表版本前，可以撤销最近一次行写入。
- 删除数据表、数据行或索引前都需要明确确认。
当前范围
当前版本仅支持本地 LanceDB，暂不包含 S3、MinIO、Embedding Provider、混合检索、Reranker、云端账号以及导入导出。
作者与许可证
| |
