Mongrel is a cross-platform database workbench that unifies 30+ database engines with terminals, file transfer, containers, Kubernetes and API testing in one native desktop app, so developers and DBAs can follow a problem across systems without switching tools.
Mongrel is a desktop database workbench that brings querying, administration and the surrounding infrastructure into a single application. You start by creating a connection profile for a supported engine, then use the shared query editor to write SQL or an engine dialect, browse schemas, edit rows inline and inspect execution plans. From the same session you can open a remote terminal, move files over SFTP or S3, inspect a running container, check a Kubernetes pod, or send an HTTP request. Saved queries, snippets, notebooks and runbooks keep repeatable investigation steps close at hand, while connection profiles can be versioned in Git so teammates share a consistent setup. The goal is simple: fewer windows, less copy-pasting, and context that survives each step of debugging.
Mongrel is built for polyglot environments rather than a single vendor stack. Compatible drivers cover relational systems, document stores, key-value engines and analytics warehouses, and each supported engine gets its own query and administration workflow inside a common connection list, inspector and results workspace. Data editing, schema comparison, migrations, import and export, and scheduled encrypted backups appear wherever the driver exposes them. Because the app is local-first, query text, results and logs stay on your machine while credentials live in the operating system keychain — the same preference for focused, self-contained desktop utilities that has kept tools such as FolderSizes useful for years. AI assistance stays off until you configure a provider, and any generated write is approval-gated.
The clearest benefit of Mongrel is context that survives a debugging session. When a slow endpoint turns out to be a missing index, you can move from the query plan to the application logs in a terminal, then into the container running the service, without reconnecting or re-authenticating in three separate programs. That continuity shortens investigations and reduces the copy-paste errors that creep in when results, hostnames and credentials are juggled between windows. A single connection list, inspector and results workspace also lowers the learning curve for polyglot teams: one set of keyboard shortcuts, one searchable catalog of saved queries, and one place where schemas, drift and metadata are visible. Versioned connection profiles in Git help teams standardize environments and review changes, while local-first storage keeps query text, results and logs on your own machine. Write gates, production guards and local audit records add a safety layer for production work, and the opt-in assistant runs only after you configure a provider. Every plan includes the full feature set, so nothing essential sits behind a higher tier.
Mongrel keeps every database, cluster and endpoint profile in one searchable connection list. Profiles can be versioned in a local Git repository with optional push and pull to a remote, in-app conflict resolution, and proposer or approver review steps, while credentials are stripped from sync payloads by default.
Write native SQL or engine-specific dialects in a shared editor that offers snippets, parameterized templates, saved query files, query history and versioned drafts. Autosave, session restoration and crash recovery mean a long-running investigation survives an unexpected restart or an edited file on disk.
Build SELECT statements with joins, filters, grouping and ordering visually for MySQL and MariaDB, then review the generated SQL before sending it to a query tab. The same workspace lets you design or reverse-engineer schemas on an ER canvas wherever drivers expose table relationships.
Edit rows directly in result grids instead of writing UPDATE statements by hand, and follow foreign-key relationships up to five hops without leaving the grid. Tracing how a value reached a table becomes far faster than assembling ad-hoc joins and switching between editors.
Open SSH, Mosh, Telnet and serial sessions beside your queries, with session logging, triggers, quick commands, guarded broadcast, themes, search and recording for replay. Server-side checks stay in the same context as the database work, so you are not hunting for the right host or credentials in another window.
Move files over FTP, FTPS, SFTP, SCP, S3, WebDAV, Azure Blob and SMB in a dual-pane browser with multi-select, resume, folder sync, drag and drop and remembered folders. A durable transfers drawer keeps queued work moving when a connection drops or a session is restarted.
Manage Docker and Podman images, containers, volumes and networks from the same session, run kubectl tasks such as log tailing, exec, port-forwarding and Helm releases, then exercise REST, GraphQL, WebSocket, gRPC, SOAP and JSON-RPC endpoints with OAuth PKCE, mTLS and imports from cURL, Postman or OpenAPI.
Mongrel ships with AI disabled until you connect an OpenAI-compatible, Anthropic or Gemini provider of your choice. API keys stay in the operating system keychain, generated queries require review before execution, and any per-profile automation you permit is narrowly read-only by default.
Mongrel is a cross-platform desktop database workbench from VisorCraft. It combines clients for more than 30 database engines with remote terminals, file transfer, container and Kubernetes management, and an HTTP and GraphQL API client in one native application, so an investigation can move between systems without switching products.
Mongrel supports 30 or more engines, spanning relational systems such as PostgreSQL, MySQL, MariaDB, SQL Server and Oracle, document and key-value stores including MongoDB and Redis, and analytics warehouses such as ClickHouse, Snowflake and BigQuery. Each engine gets query and administration workflows wherever its driver exposes them.
No. AI is disabled until you configure a provider, and providers are entirely optional. When enabled, API keys are stored in the operating system keychain, generated queries are treated as drafts that need review, and per-profile automation stays read-only unless you deliberately widen it.
Mongrel is local-first: query text, results, application data and logs stay on your machine and are not sent to VisorCraft automatically. Network requests happen only for services you configure, such as licensing, update checks, plugin catalogs, or an AI provider you have deliberately connected.
Yes. Connection profiles can be versioned in a local Git repository, with optional push and pull to a remote, in-app conflict resolution, and proposer or approver review steps. Secrets are kept out of normal profile exports by default, while private-key and known-hosts paths stay machine-local.
Yes. Connections can be direct or TLS-encrypted, and per-connection SSH tunneling is available where the driver supports it. Passwords and secret values are stored using the operating system credential facility when one is present, so they are not kept in plain profile files.
Mongrel is a native desktop application available for Windows, macOS and Linux, so the same workbench, feature set and licensing flow is available whichever system you run it on.
Yes. A seven-day free trial is issued once per machine through the licensing service and does not require a credit card. Because every plan includes the complete feature set, the trial gives you an accurate picture of what the paid product does.
Comments