SQLly vs Peek
Peek is a free macOS database client built with Rust and Tauri that puts queries and results on an infinite canvas, Figma-style. Click a foreign key and the related rows open as a new node; host a session and a teammate joins peer to peer. It speaks PostgreSQL, MySQL and MariaDB today and is a public beta. SQLly is a paid, closed-source workbench that goes wider on engines and deeper on SQL, guardrails, and DBA tools. Here’s the honest breakdown.
Last reviewed: November 2026
The short version
A deep SQL workbench
One Rust codebase drawn on the GPU on macOS, Windows, and Linux, with no browser engine. Thirty engines, IntelliSense that reads your data, layered safety guardrails, schema compare, and a full set of script-first admin tools.
Your queries on a canvas
A Tauri 2 app: a Rust backend (sqlx, an embedded tree-sitter SQL language server, iroh for multiplayer) with a React and Monaco front end. Query, result, chart, agent, variable, text and drawing nodes on one canvas, foreign keys you can click through, and live P2P sessions. Version 4.4.0 for Apple Silicon Macs, built mostly by one developer.
Feature by feature
SQLly’s column uses the same honest status badges as the features page and the docs. For Peek we list what its website, docs, changelog, release assets and source state.
| Capability | SQLly | Peek |
|---|---|---|
| Runs on macOS, Windows & Linux | YesmacOS is signed and notarized; Windows and Linux are a supported beta, on amd64 and arm64. | macOS onlyA signed, notarized .dmg for Apple Silicon. No Windows or Linux builds are published; you can build from source with Rust, Node and Yarn. |
| Intel Macs | NoApple Silicon only. | NoEvery release ships an aarch64 DMG only. |
| How the app is built | NativeOne Rust codebase, rendered on the GPU (Metal, Direct3D, Vulkan). How it works | TauriRust backend (sqlx, russh, iroh); React and Monaco front end in the system webview. |
| Download size | LargermacOS dmg 65.6 MB, Windows x64 zip 60.9 MB, Linux x64 tarball 76.2 MB. | Smallerv4.4.0: Apple Silicon dmg 25.3 MB. |
| Free to use | Free tierTwo saved servers, one cloud account, and a reminder to pay about every five hours. Pricing | YesA free download; no price or paid tier on its site or repository. |
| Open source | NoLifetime licenses carry a source-escrow promise if SQLly is ever discontinued. | Source publicThe code is on GitHub and its maker calls it open source, but the repository has no license file. |
| PostgreSQL, MySQL & MariaDB | YesNative drivers. Full list | YesThrough sqlx. MySQL and MariaDB reached parity with Postgres in 4.4.0 (September 2026). One live connection at a time. |
| SQLite & DuckDB | YesDuckDB is bundled. Per-engine coverage | NoIts homepage lists both, but the 4.4.0 source has drivers only for postgres:// and mysql:///mariadb:// URLs and rejects anything else. |
| SQL Server, Oracle, ClickHouse, libSQL/Turso | YesPlus Azure SQL and wire-compatibles such as Neon, Supabase and Aurora. Compatibles | No |
| Redis | YesTyped key editing, Cluster and Sentinel, pub/sub, profiler, memory analyzer; Valkey, KeyDB, DragonflyDB, ElastiCache and MemoryDB too. Redis | No |
| Infinite canvas workspace | NoTabs, split editors and a results grid. | YesQuery, result, chart, agent, variable, text and freehand drawing nodes, with pages, regions, a minimap and per-page version history. |
| Follow foreign keys | PartialThe Relationship Graph maps the keys and IntelliSense writes JOIN ON from them, but there is no one-key jump from a cell to the referenced row. |
YesClick a foreign key to open the related rows as a new node; click a primary key to see every row that references it. |
| Live collaboration | NoConnections and settings are plain files, exportable without passwords. Everything is a file | YesPeer-to-peer sessions over iroh with live cursors. Guests, including browser guests, see only the connection name; their queries run on the host’s connection. |
| SSH tunnels & TLS | YesSSH hops, SOCKS and HTTP proxies, and TLS modes. TLS | PartialOne SSH tunnel per connection, by key file or password. Its source accepts any SSH host key without checking it. |
| Saved passwords | OS keychainConnection files are plain text; passwords live in the keychain. | Plain fileIts docs put the password inside the connection URL in a workspace JSON file under ~/peek. We found no keychain integration. |
| IntelliSense | YesAlias- and scope-aware, JOIN ON from foreign keys, real column values in WHERE. IntelliSense |
YesAn embedded tree-sitter language server: alias-aware tables and columns, JOIN ON snippets from foreign keys, and squiggles for unknown tables and columns. |
| Edit table data | YesA row editor generated from the table’s own metadata. Row editor | YesDouble-click a cell to edit it; the schema picks a dropdown or a JSON editor where it fits. |
| Query history | YesSearchable, with pinned snippets. Query History | YesPer-connection history, opened from the command palette. |
| Large results | YesStreams and spills to disk past 50,000 rows or 64 MiB; flat memory at 50 million rows. Streaming | -Result tables are virtualized and its site cites 10k rows; we couldn’t confirm how much bigger results are held. |
| Charts & pivot | YesPivot, charts, conditional formatting, computed columns, and diffs. Pivot | ChartsChart nodes for numeric results, and results side by side on the canvas. No pivot that we could find. |
| Import & export data | YesCSV, JSON, SQL, and Excel with live formulas and a pivot sheet. Export | PartialExport CSV, JSON and SQL INSERTs. Drop CSV, JSON or Parquet files onto a connection to query them as temporary tables. |
| Monitor & kill running queries | YesKill is scripted for review, never run on click. Activity Monitor | YesA Running queries node (4.3.0) with a click-to-kill button per query. |
| Execution plan viewer | In progressPlain-language findings, tree, and diagram across nine engines. Query plans | -Not on its site, docs or changelog. |
| ER diagrams | YesExport to SVG, Mermaid, or DBML. Relationship Graph | YesA schema graph view. |
| Schema compare, designers, backup | YesRisk-graded schema diffs, table designers, and backup and restore as reviewable scripts. Tools | NoNone on its site or in its docs. |
| Guardrails in the editor | YesSELECT-only connections, no UPDATE/DELETE without WHERE, auto-rollback, tenant checks, prod in red. Safety |
PartialA confirmation before DELETE without WHERE or TRUNCATE, and color-coded connections. Read-only connections are an open issue. |
| AI assistant & MCP server | YesLocal models, your own endpoint, or opt-in cloud; a read-only MCP server. How reads are kept read-only | YesAgent nodes on your own OpenAI-compatible endpoint (Ollama, llama.cpp, LM Studio) or Claude Code over ACP. The MCP server is off by default; when on, its docs say it listens on all interfaces with no authentication. |
| Command palette & themes | YesThe palette covers the whole menu bar; twelve themes in light and dark; rebindable keys. Themes | YesA keyboard-first palette for commands, pages, queries and tables; three themes; a rebindable keymap. |
A dash means we couldn’t confirm it from Peek’s own materials - check its repository for the current state.
Where SQLly is different
Peek rethinks how you look at query results. SQLly stays with tabs and a grid, and spends its effort on engines, the SQL, and the work around it.
Windows and Linux too
SQLly runs from the same Rust codebase on macOS, Windows, and Linux, on amd64 and arm64. Peek publishes Apple Silicon builds only.
Cross-platform →Thirty engines, Redis included
SQL Server, SQLite, DuckDB, ClickHouse, Oracle, libSQL/Turso and a full Redis toolset sit next to the PostgreSQL, MySQL and MariaDB that both apps share.
Databases →Guardrails beyond a confirm box
Read-only connections that refuse file-reading and shell-escape tricks, a hard stop on UPDATE/DELETE with no WHERE, auto-wrapped transactions, and prod painted red - inherited down your client, project, and environment tree.
Passwords in the keychain
SQLly’s connections are plain files you can sync or check into git, but passwords stay in the OS keychain and never leave it through import or export.
File-based sync →An MCP server that can only read
An agent gets schemas and a single SELECT in a read-only transaction that is always rolled back; anything that writes comes back to you as a proposal.
Compare, transfer, and admin tools
Schema compare across engines with sync and rollback scripts, data compare, cross-engine data transfer with type mapping, SQL Agent, Extended Events, an index analyzer, and a test data generator.
Tools →When Peek is the better choice
Peek is probably the better fit if:
- You explore data by following relationships. Clicking from an order to its customer to their other orders, with every step left on the canvas, is what Peek is built around. SQLly has no one-click jump from a cell to the referenced row.
- You want to look at data together, live. Peek’s peer-to-peer sessions give a teammate your canvas and cursors, even in a browser, without handing over credentials. SQLly has nothing like it.
- You think better on a whiteboard. Laying queries, results, charts, notes and sketches out side by side suits investigations and explaining a schema to someone.
- You want an agent to build the board. Peek can hand its canvas to Claude Code over ACP or MCP, so the agent writes queries and arranges the nodes.
- You only use PostgreSQL or MySQL on an Apple Silicon Mac and want a free, 25 MB app. SQLly’s free tier stops at two saved servers and isn’t available inside organizations with more than 100 employees.
One honest caveat in the other direction: Peek is a public beta at version 4.4.0 (September 2026), nearly all of its commits come from one developer, and its repository has no license file, so check its repository for recent activity before you rely on it.
Questions
Is Peek free?
Peek is a free download, and its site and repository list no price. Its source is public on GitHub, but the repository has no license file, so check with its maker before you reuse the code. SQLly is free for hobbyists (two saved servers), with a one-time $100 SQLly 1.x license and a $200 Lifetime license for everyone else.
Which databases do SQLly and Peek support?
Peek 4.4.0 connects to PostgreSQL, MySQL and MariaDB, one connection at a time. Its website also lists SQLite and DuckDB, but we found no driver for either in its source, which rejects other connection URLs. SQLly connects to all three plus SQL Server, SQLite, DuckDB, ClickHouse, Oracle, libSQL/Turso and Redis, and documented wire-compatible services such as Neon, Supabase and Aurora - 30 engines in all.
Does Peek have AI or an MCP server?
Yes. Peek’s agent node runs against your own OpenAI-compatible endpoint such as Ollama, or an external agent such as Claude Code over ACP, and its optional MCP server lets agents create and edit queries on the canvas. The MCP server is off by default; its docs say that when enabled it listens on all network interfaces with no authentication. SQLly works with local models (Ollama, MLX), your own endpoint, or opt-in cloud providers, and its sqlly-cli mcp lets an agent read schemas and run a single SELECT in a read-only transaction that is always rolled back; writes come back to you as proposals.
Other comparisons
Point both at the same database
The quickest way to choose is to run your own queries in each. SQLly is free on macOS, Windows, and Linux, with no account required.