QueryLane

English

Русский

English

Русский

Appearance

Download

QueryLane for Mac, Windows and Linux

All your data,
one place to work with it

Query any database, turn results into jobs, charts and dashboards, and bring in an AI assistant whenever it helps

Download for MacRead the docs

Free · Apple silicon and Intel · Signed and notarized by Apple

Sources

Works with the databases you already use

Connect relational, document and analytics databases, query CSV, JSON and Parquet files on your computer, in S3 or over SFTP with SQL, and call HTTP APIs from your jobs. Everything opens in the same explorer, editor and result view

PostgreSQLRelational
MySQLRelational
MariaDBRelational
PerconaRelational
CockroachDBDistributed SQL
YugabyteDBDistributed SQL
SQL ServerRelational
MongoDBDocument
ClickHouseAnalytics
SQLiteEmbedded
CSVCSV · JSONQuery with SQL
HTTP APIInside jobs
More through the JDBC pluginInstall it in Settings, and these databases open in the same explorer, editor and jobsHow it works
OracleEnterprise
SnowflakeCloud warehouse
DatabricksLakehouse
IBM Db2Enterprise
BigQueryCloud warehouse
TeradataEnterprise
Apache HiveAnalytics
Spark SQLAnalytics
TrinoQuery engine
FirebirdRelational
H2Embedded
Any JDBC driverYour own driver

SSH tunnels are built in, and SQL Server works from 2017 on, with SQL Server or Windows logins

Product names and logos belong to their owners. Firebird is a registered trademark of the Firebird Foundation Incorporated

AI, if you want it

Work your way, with or without AI

Bring in the assistant when it saves you time. It knows your schema, writes queries and builds steps, and it always asks before running anything or making changes

By handClickHouse · Event stream
SELECT
    toStartOfWeek(ts) AS week,
    app_version,
    uniqExact(user_id) AS active_users
  FROM events
  WHERE ts >= now() - INTERVAL 8 WEEK
  GROUP BY week, app_version
  ORDER BY week
With the assistantSame result
Show weekly active users by app version for the last 8 weeks. Events are in ClickHouse
I checked the events table and wrote the query. It returns 24 rows, 8 weeks for each of 3 versions
Active users by versionSQL query · Event stream · 24 rows
Add the step to “Weekly product telemetry”?Allow onceAllow in this chatDeny
Ask about your datalocal · Ollama
ModelsChoose OpenAI, Anthropic, Google or OpenRouter, or run a local model with Ollama, LM Studio or any OpenAI-compatible server
What the model seesShare only the schema, anonymized rows or real data, whichever fits the task
Your own agent over MCPConnect Claude Code, Claude Desktop, Codex, Cursor, VS Code and others in one click, with the access you set for each connection

MCP

Bring the AI agent you already use

QueryLane has its own MCP server. Your agent reads schemas, builds and runs jobs and makes dashboards with its own model and account, and QueryLane decides what it may touch

Set up in one clickClaude CodeClaude DesktopCodexCursorVS CodeZ CodeOpenCodeHermes AgentOpenClawAny MCP client
Ask in your agentYour agent · MCP
Build a weekly dashboard of new signups by country from our Postgres
querylane_schema_getread the schemaquerylane_job_createmade the jobquerylane_run_start12 rowsquerylane_dashboard_create3 widgets
Done. The job runs every Monday at 08:00, and the dashboard reads its results
Signups by countryDashboard · 3 widgetsOpen in QueryLane
Reply to your agentMCP · querylane
You stay in chargePer connection
  • Access levelsNo access, read only, or read and write for each connection. Production starts read only
  • ApprovalsChanges on a connection set to Ask wait for your Allow in the app
  • No passwords for agentsSaved passwords stay in the encrypted vault and never reach the agent or another server
  • Your files stay yoursAgents save files only into their own folder and ask before reading or uploading yours
  • Action journalEvery call with the client that made it and its result

The agent works on its own model and subscription, QueryLane needs no AI account for it How to connect an agent

Product names belong to their owners. QueryLane is an independent product and is not affiliated with or endorsed by them

Jobs

From a question to a pipeline

Combine queries, JavaScript, HTTP requests, files and plugins in a single job, load the results into any table or storage, add checks that keep bad data out, and run it right away or on a schedule

AccountsPostgreSQL
App eventsClickHouse
DevicesMongoDB
Support ticketsMySQL
Feature flagsHTTP
JSJoin and groupJavaScriptcheck
Chartin the app
Telegramevery morning
CSV exporton demand
Report tableevery night
Steps that depend on each other

Each step waits for the results it needs, and independent steps run in parallel

Checks that catch bad data

A failed check can add a warning or stop the job before the data goes anywhere

Schedules that run on their own

Jobs run on time even when QueryLane is closed. You hear about failures right away, and Telegram can deliver every result

Jobs in git

Keep jobs as YAML in a repository and clone, pull, commit and push them with full file history

Version history

Every step change is saved with a diff, so you can restore any version, even a deleted step

Packages to share

Send a job as a single file. Secrets stay behind, and connections are matched on import

See it run

From the graph to a chart in one run

Open a job, run it and the report is ready a moment later. Every step runs on your computer

Dashboards

Reports that keep themselves fresh

Put numbers, tables and charts from your jobs on one page. A dashboard opens at once from stored results and picks up scheduled runs, so the morning report is ready before you are

From a job in one click

Add a chart from a run report, or make a dashboard of every chart a job draws

Filters without reruns

Narrow every widget by dates, values or ranges right over the stored rows

Parameters

Change a job variable and the dashboard runs the job in the background with it

Always up to date

Show the latest runs, run the jobs on open or follow scheduled runs live

Save and share

Save as an HTML page or an image, or export with its jobs to another QueryLane

Built for you by AI

Ask the assistant or your MCP agent for a report and it builds the jobs and the dashboard

In practice

What people build with QueryLane

AccountsApp eventsDevicesTicketsFlagsJSActive usersby app version8 weeks · 24 rows

Telemetry from five databases

A fintech team brought events, accounts, devices and tickets together in one weekly chart

Morning reportRuns daily at 08:00Send to TelegramPlugin step · sent in 0.4 s8:00QueryLane ReportsbotMorning reportThu, Sep 25Active users60.1kNew accounts1,284Card declines3.2%Open tickets4108:00

A morning report in Telegram

A scheduled job sends the summary before the workday begins

card_declines.sqlResultExplain123456SELECT m.country, count(*) AS declinesFROM card_events eJOIN merchants m ON m.id = e.merchant_idWHERE e.status = 'declined'AND e.created_at > now() - interval '30 days'GROUP BY 1 ORDER BY 2 DESC;Explain1.9M rows scanned · 1.8 sSortHashAggregateHash JoinIndex Scan · card_events_status_idxSeq Scan · merchants64641.9M1.9M12k

Heavy SQL on a large database

Autocomplete, explain plans and paging through millions of rows

Code

When SQL isn’t enough, reach for JavaScript or Python

Any step can be code. Use earlier results by name and hand rows to the next step

  • npm packages added in Settings
  • Python steps with pandas, Telegram messages and a headless browser through plugins
  • Screenshots and files from plugin steps, saved with each run
// Active users by version, split by plan
  const events = ${app_events};
  const plans = new Map(
    ${accounts}.map(a => [a.user_id, a.plan])
  );

  const groups = {};
  for (const e of events) {
    const plan = plans.get(e.user_id) ?? 'free';
    const key = `${e.week}|${e.app_version}|${plan}`;
    groups[key] = (groups[key] ?? 0) + 1;
  }

  return Object.entries(groups).map(([key, users]) => {
    const [week, version, plan] = key.split('|');
    return { week, version, plan, users };
  });

Database client

A database client you’ll enjoy opening

Browse, query, edit and export data in any engine through one calm interface

Explorer

Schemas, tables, views, indexes, functions, collections and GridFS, all in one tree

Query editor

Autocomplete from your schema, saved queries, parameters and explain plans on every engine

Table, JSON or tree

Switch how any result looks without running the query again

Edit in place

Change rows right in the grid, preview schema changes as DDL and commit or roll back transactions

Connect in seconds

Paste a connection string, or bring your connections over from DBeaver and Navicat with their passwords

Import and export

Export to CSV, JSON or Excel, import CSV, XML or Excel files, copy data between connections and open S3 or SFTP files as tables

Your data stays
on your computer

Connections, queries and results live on your machine and reach an AI provider only when you connect one. With a local model, your data never leaves your computer

Passwords guarded by the system

Passwords, API keys and tokens sit in an encrypted vault. Its key lives in the Keychain on a Mac, is protected by Windows for your account on a PC and sits in the system keyring on Linux, and AI agents only ever see masks

Access you choose for agents

For each connection, agents get no access, read only or read and write. Read only also opens a read-only database session wherever the engine has one

Code on your terms

One switch stops agents from running JavaScript, HTTP and plugin steps on your computer

Changes wait for you

Set a connection to Ask, and every change an agent makes waits for your answer in QueryLane. The action journal shows what agents did

Roadmap

What’s next

In order of priority

  1. Safer agents

    Allowed schemas and undo on top of the access per connection, approvals and action journal that are already here

  2. Self-hosted

    A server of your own that runs published jobs with retries, backfills and dependencies, also while your computer is off

  3. Cloud

    The same runner hosted by us, for teams that would rather not keep a server

Pricing

The desktop app is free

The desktop features are free to use on Mac, Windows and Linux. Cloud and self-hosted options are planned

Desktop

Available now
Free
  • Every database, file and HTTP source
  • Jobs, checks, charts, dashboards and schedules
  • AI assistant with your own keys or local models
  • MCP for your own agents, plugins
Download for Mac

Cloud

Planned
Coming soon
  • Jobs that keep running while your computer is off
  • Orchestration with history and retries
  • Cloud runs to test an idea quickly
  • Pipelines stored and shared in the cloud

Details and availability will appear on this site

Self-hosted

Planned
Coming soon
  • Run QueryLane on infrastructure you manage
  • Keep execution and data in your environment

Deployment details and availability will appear on this site