WhoDB

An honest look at the AI-powered open-source database management tool. What you actually get when you self-host it.

TL;DR

• What it is: Open-source (Apache-2.0) database management and exploration tool built with Go and React — think a lighter, AI-capable DBeaver that doesn’t require Java [README].
• Who it’s for: Developers and technical founders who bounce between multiple database types (PostgreSQL, MySQL, MongoDB, Redis, SQLite, ClickHouse, and more) and want one clean UI instead of five separate tools [README][1].
• Cost savings: Traditional database GUIs range from $69 (TablePlus, one-time) to $129/year (DataGrip). WhoDB is Apache-2.0 — free to self-host, no per-seat licensing.
• Key strength: Genuinely lightweight (<50MB Docker image), instant startup, and a natural language chat interface that works with local models via Ollama — no OpenAI API key required [README].
• Key weakness: Small project (4,679 GitHub stars as of this review), thin third-party coverage, and the “Fortune 500” and “SOC 2/GDPR/HIPAA compliant” marketing claims on the homepage are unverified by any independent source. Treat the enterprise claims with skepticism until the community has more time to stress-test them.

---

What is WhoDB

WhoDB is a database management and exploration GUI. You point it at a database — PostgreSQL, MySQL, SQLite, MongoDB, Redis, MariaDB, Elasticsearch, ClickHouse, Cassandra, MSSQL, or others from a claimed list of 40+ — and get a web interface for browsing tables, running queries, visualizing schema relationships, and optionally talking to your data in plain English [README].

The pitch in the GitHub README is blunt: “A lightweight next-gen data explorer — Postgres, MySQL, SQLite, MongoDB, Redis, MariaDB, Elastic Search, and Clickhouse with Chat interface.” That’s an accurate one-liner. The homepage layers “enterprise-grade” and “Fortune 500” on top of it, which is standard startup marketing inflation you can mostly ignore when evaluating the software itself.

What’s technically real: the core is a Go backend with a React frontend. The Docker image is under 50MB. Startup is fast enough that the README claims under one second. The AI layer lets you type “show me all users who signed up last week” instead of writing SQL, and it supports Ollama (fully local inference), OpenAI, Anthropic, and any OpenAI-compatible API endpoint [README].

The project was first spotted publicly in late June 2024 when selfh.st’s weekly self-hosted newsletter flagged it as a newly launched “user-friendly database management tool” [1]. Since then it has accumulated 4,679 GitHub stars — respectable but not dominant for the category.

---

Why people choose it

Third-party reviews are thin. The project is young enough that dedicated comparison articles don’t yet exist. What we have are GitHub activity signals, the selfh.st community pickup [1], and the README’s own feature narrative. Synthesizing these:

The DBeaver alternative angle. DBeaver is the category default for multi-database management — free, open source, handles almost every database, and weighs about 200MB with a Java runtime. It also has a UI that looks like it was designed in 2012 by someone who prioritized options over clarity. WhoDB positions itself as the modern replacement: same multi-database breadth, fraction of the resource footprint, and a design that looks like it was built in the last three years.

The “no more switching tools” argument. Most development setups involve multiple databases at once. PostgreSQL for the app, Redis for caching, MongoDB or Elasticsearch for something specific. Every tool that only handles one type is friction. WhoDB’s explicit goal is a single interface for the whole stack [README].

The local AI angle. The most interesting differentiator is the Ollama integration. You can run natural language queries against your production PostgreSQL without sending any data to OpenAI — the inference happens on your machine or your own server [README]. For founders handling sensitive user data, that’s not a small thing. You get the “talk to your database” UX without the data-leaving-your-server problem.

Community reception. The selfh.st newsletter is a curated list — inclusion in a June 2024 weekly as a notable new launch signals it passed a basic quality filter [1]. The GitHub commit activity badges in the README show monthly commits and recent active development. That said, 4,679 stars is still early-stage. There’s no equivalent of Activepieces’ 21,000 stars or n8n’s 100K+ here. The project is genuinely small.

---

Features

Based on the README and website scrape:

Database support:

• PostgreSQL, MySQL, SQLite, MariaDB [README]
• MongoDB, Redis [README]
• Elasticsearch, ClickHouse [README]
• Cassandra, MSSQL [features list]
• Website claims 40+ total, including Oracle and DynamoDB — the full list isn’t in the README

Data management:

• Spreadsheet-like data grid with inline editing
• Sort, filter, search within tables
• Bulk operations
• Query history and management
• Mock data generation for dev environments [README]
• Flexible export options

Schema and relationships:

• Interactive graph visualization of table relationships
• Foreign key navigation
• Schema topology explorer with pan, zoom [README]

AI features:

• Natural language to SQL — type a question, get a query [README]
• Conversational interface (“Chat interface” is in the GitHub description)
• Supports Ollama (local, no API key), OpenAI, Anthropic, and any OpenAI-compatible endpoint [README]
• Works fully offline with Ollama — no data leaves your network

Deployment options:

• Docker (primary)
• Docker Compose
• Helm / Kubernetes [README]
• Windows app (Microsoft Store)
• macOS app (App Store)
• Snap (Linux)
• CLI
• npm

The cross-platform distribution is broader than most tools in this category. Docker is the expected path for self-hosters, but having native Windows and macOS apps means it works equally well as a local desktop tool or a server-hosted team resource.

---

Pricing: SaaS vs self-hosted math

No SaaS pricing tier is documented in any of the sources reviewed. The website describes itself as self-hosted and open source (Apache-2.0), with no visible subscription or cloud-hosted offering [README][website]. The enterprise marketing language on the homepage suggests a commercial tier may exist or be planned, but pricing is not available from any reviewed source.

Self-hosted cost:

• WhoDB itself: $0 (Apache-2.0 license)
• Server to run it: $5–10/month on any VPS if you want it always-on
• For local use: $0 — run it on your laptop, point it at a local or remote database

What you’d otherwise pay for comparable tools:

Tool	Model	Cost
DataGrip (JetBrains)	Annual subscription	$129/year (~$10.75/mo)
TablePlus	One-time + upgrade	$69 one-time
DBeaver Community	Free	$0
Beekeeper Studio Community	Free (limited)	$0
WhoDB	Free (Apache-2.0)	$0

The honest comparison is against free tools: DBeaver and Beekeeper Studio Community. If you’re already using one of those and it works, WhoDB’s cost argument doesn’t apply — the pitch is UX and the AI layer, not price.

If you’re currently paying for DataGrip or TablePlus and spending most of your time in CRUD operations and schema exploration rather than advanced IDE features, the math is obvious.

---

Deployment reality check

What you need for Docker deployment:

• Docker installed
• 512MB–1GB RAM (lightweight is real — the <50MB image claim is for the container image, runtime footprint will vary with query size and connections)
• A database to connect to (WhoDB doesn’t host databases — it’s a client)
• A domain and reverse proxy (Caddy or nginx) if you want HTTPS access from outside localhost

The quick-start path from the README:

docker run -it -p 8080:8080 clidey/whodb

That’s it for local exploration. No configuration file, no environment setup, no database to provision. You add connections through the UI after launch. For a developer laptop tool, this is one of the faster setups in the category.

What can complicate things:

• The AI features require you to either have an API key (OpenAI/Anthropic) or run Ollama separately. WhoDB doesn’t bundle a local model — you set that up yourself, which is a non-trivial process for non-technical users [README].
• The “40+ databases” claim includes Oracle and DynamoDB, which have their own driver/connectivity requirements that may not be as turnkey as PostgreSQL.
• No third-party setup guides exist yet (project is young). If you hit an edge case, you’re relying on GitHub issues and the project’s own documentation.

On the enterprise claims: The homepage’s “SOC 2/GDPR/HIPAA compliant” language is unusual for a self-hosted open-source tool. Compliance is a property of your deployment and data handling practices, not the software itself. No third-party audit report is linked. No independent source confirms these claims [website]. Treat this as marketing aspiration, not verified certification.

---

Pros and cons

Pros

• Apache-2.0 license. Genuinely permissive — use it commercially, embed it, fork it, no Fair-code restrictions [README].
• Actually lightweight. Under 50MB Docker image and sub-second startup aren’t marketing numbers — Go binaries are small, and the README’s performance table claims 90% less resource usage than “traditional tools” [README]. Even if that’s cherry-picked, it’s in the right direction.
• Multi-database, one UI. Postgres + MongoDB + Redis in a single client is the actual use case, and it handles it [README].
• Local AI that doesn’t phone home. Ollama integration means natural language queries work entirely offline. For any database holding sensitive data, this is the right architecture [README].
• Native apps on all platforms. Docker, Windows Store, Mac App Store, Snap, CLI — distribution is thorough for a small project [README].
• Modern UI. The screenshots in the README look clean and contemporary. Schema topology visualization looks genuinely useful rather than decorative.
• Active development. Monthly commit badges and recent release dates suggest the project isn’t stalled [README].

Cons

• Small community, limited track record. 4,679 stars and nearly no third-party reviews means you’re early. Bugs may take longer to surface and fix. If you need mission-critical tooling with a support track record, this isn’t there yet [1].
• AI setup requires external work. The chat interface is the headline feature, but getting Ollama running is a separate project. Non-technical users can’t just click and get AI queries [README].
• Unverified enterprise claims. “SOC 2/GDPR/HIPAA compliant” and “Fortune 500 trusted” appear on the homepage with no supporting evidence in any reviewed source [website]. Don’t buy a vendor’s compliance story without an audit report.
• Limited documentation. Docs live at docs.whodb.com, but there’s no community of tutorials, YouTube walkthroughs, or third-party setup guides yet. You’re somewhat on your own.
• No built-in authentication beyond the connection screen. The README doesn’t describe user management, RBAC, or team access controls. Suitable for personal use or small trusted teams; not ready for organization-wide access management without additional infrastructure.
• Niche database support claims unverified. PostgreSQL and MySQL coverage is likely solid. Oracle and DynamoDB at the edge of the “40+ databases” list — no review has tested these.

---

Who should use this / who shouldn’t

Use WhoDB if:

• You work across multiple database types daily and are tired of switching between DBeaver, Redis Insight, MongoDB Compass, and a separate Elasticsearch GUI.
• You want natural language queries against your own data without sending queries or results to external AI APIs (Ollama).
• You want a database client that starts in under a second and doesn’t require a JVM.
• You’re a solo developer or small team — the missing team access controls don’t matter at that scale.
• You’re willing to be slightly early on a project and can tolerate rougher edges.

Skip it (use DBeaver instead) if:

• You need proven, battle-tested multi-database support with years of community fixes and plugins.
• You rely on advanced IDE-style features: ER diagrams with export, migration tools, CSV import wizards.
• You need your team to all use the same tool with verifiable setup docs.

Skip it (use Beekeeper Studio Community instead) if:

• You primarily work with SQL databases (PostgreSQL, MySQL, SQLite) and want a clean free GUI without the “40 databases” overhead.
• You value a more established open-source community with a longer history.

Skip it (use DataGrip instead) if:

• You’re a JetBrains shop, already paying for other JetBrains tools, and want deep IDE integration with your editor.
• You need advanced refactoring, schema migration tooling, and query plan analysis.

Skip it if:

• Your compliance team requires certified vendor documentation — the homepage claims aren’t backed by anything reviewable.
• You’re a non-technical founder who wants a GUI with zero setup. The Ollama requirement alone adds meaningful setup complexity for anyone without a terminal background.

---

Alternatives worth considering

• DBeaver — the incumbent open-source option. Free Community edition, every database imaginable, Java-heavy, UI shows its age. Still the default recommendation for broad database support.
• Beekeeper Studio — MIT-licensed community edition, notably cleaner UI than DBeaver, primarily SQL databases. If you’re SQL-only, this is often the better pick.
• TablePlus — paid ($69+ one-time), native apps, widely praised UI, strong PostgreSQL and MySQL support. The design reference point WhoDB seems to be aiming at.
• DataGrip — JetBrains’ database IDE, $129/year, most powerful feature set in the category, overkill for simple exploration.
• Adminer — single PHP file, free, handles MySQL/PostgreSQL/SQLite/MongoDB. Ugly but portable. Use it when you need quick access on a server that already runs PHP.
• pgAdmin — PostgreSQL-specific, free, the default admin tool for Postgres. Messy UI but the most feature-complete for Postgres administration.
• Redis Insight — Redis-specific, free, made by Redis Ltd. If Redis is your main concern, this beats any generic multi-DB tool.

For a developer who touches PostgreSQL, Redis, and MongoDB regularly and wants one tool: WhoDB versus DBeaver is the real decision. DBeaver wins on stability and community. WhoDB wins on resource footprint, UI quality, and AI features. The tradeoff is clear.

---

Bottom line

WhoDB is the right bet for developers who are fed up with running three separate database clients or paying $130/year for DataGrip’s features when they mainly need table browsing and query execution. The Go/React stack delivers on the “lightweight” promise in a way Java-based tools can’t, the Ollama integration is genuinely differentiated for anyone handling sensitive data, and the multi-database scope is real. The caveats are also real: this is an early-stage project with thin third-party validation, unverified enterprise compliance claims, and a community that’s still building up. The Apache-2.0 license removes one concern. The small team behind it raises a different one. If you need the category leader with years of battle-hardening, use DBeaver. If you want the category’s best UI with local AI and you’re comfortable being an early adopter, WhoDB is worth an afternoon.

---

Sources

1. Ethan Sholly, selfh.st — “This Week in Self-Hosted (28 June 2024)” — WhoDB listed as new software launch. https://selfh.st/weekly/2024-06-28/

Primary sources:

• GitHub repository and README: https://github.com/clidey/whodb (4,679 stars, Apache-2.0 license)
• Official website: https://whodb.com
• Live demo: https://demo.whodb.com
• Documentation: https://docs.whodb.com