This is a tool I created on the side that is able to parse connection strings for various data sources including SQL, NoSQL, files, cloud storage (s3, etc), and more.

I created it because I wanted to add automatic parsing for WhoDB (day job) so that users do not have to type in each detail individually when they add a new data source.

The parser currently supports SQL, NoSQL, file-based, and cloud storage connection strings (Postgres, MySQL, MongoDB, Redis, S3, SQLite, and more).

Each datasource is defined via YAML:

- required fields

- optional fields

- credential handling

- query params

- validation rules

Usage is simple:

const result = parse('postgres://user:pass@localhost:5432/app?sslmode=require');

and the result:

{ "scheme": "postgres", "type": "database", "authority": { "host": "localhost", "port": 5432 }, "resource": { "type": "database", "name": "app" }, "path": "", "query": { "sslmode": "require" }, "fragment": null, "credentials": { "username": "user", "password": "pass" }, "options": {}, "raw": "postgres://user:pass@localhost:5432/app?sslmode=require", "safe": "postgres://user:**@localhost:5432/app?sslmode=require" }

which can they be used directly in the application.

Comments URL: https://news.ycombinator.com/item?id=48194799

Points: 3

# Comments: 0

WhoDB (https://github.com/clidey/whodb) is a database management tool for SQL and NoSQL databases, runs in your browser, can be launched via Docker, and is available on the Apple, Microsoft, and Snap store. The WhoDB cli is both an interactive and command driven client that shares the same WhoDB Go backend and provides a TUI around it alongside an MCP server.

I use the mcp within Claude Code to help me when I develop with local databases (checking schemas, data, running actual SQL, etc) although CC is pretty good with this in general. I haven't yet tried it out with Codex so cannot comment on that.

We've added (in no particular order):

- DuckDB support

- TiDB support

- Updated layout system to support multiple windows

- SQL formatter and extended autocomplete

- SSH support (still testing this one)

- ER diagram generation

- EXPLAIN plan generation

- Mock data generation

- Simple table data audit functionality

- Read-only mode

- Streaming AI chat

- CLI profiles

- Different theme support

- JSON, NDJSON, CSV outputs

Disclaimer: a lot of the CLI code was AI-generated (mostly Claude Code). I point it to existing repos using the same libraries as reference, which helps since I don't have deep CLI development experience.

If you want to install it, there are several ways:

(1) curl -fsSL https://raw.githubusercontent.com/clidey/whodb/main/cli/inst... | bash

(2) brew install whodb-cli

(3) npm install -g @clidey/whodb-cli

(4) docker pull clidey/whodb-cli:latest

(5) Grab the executables from our latest Github release

If you want to add it to CC:

(1) claude mcp add whodb -- npx -y @clidey/whodb-cli mcp serve

(2) or you can configure it with database credentials via an env variable profile: claude mcp add whodb -e WHODB_POSTGRES='[{"alias":"dev","host":"localhost","user":"user","password":"pass","database":"mydb","port":"5432"}]' -- whodb-cli mcp serve

It should work with Codex as such: codex mcp add whodb -- whodb-cli mcp serve

Happy to hear your thoughts on how you might use this, or what additions you might find interesting. Thanks for reading!

Comments URL: https://news.ycombinator.com/item?id=47875513

Points: 3

# Comments: 0

How is it priced? I couldn't see anything on the site.

I am not a fan of the dbeaver, beekeeper, adminer, etc experience because they are bloated, ugly, and at best okay but not great.

Hence why I started working on WhoDB.

The approach:

- browser-based (chrome/firefox)

- no bloat

- jupyter notebook-like experience (Scratchpad)

- built-in AI co-pilot with ollama (local) or openai/anthropic

We just shipped query history and replay (time travel?) to the Scratchpad.

Would love for you to check it out and give some feedback aka roast us into oblivion:

docker run -p 8080:8080 clidey/whodb

Food for thought:

1. What's your biggest database pain point?

2. Any killer feature missing from current tools?

Comments URL: https://news.ycombinator.com/item?id=43648713

Points: 43

# Comments: 18

Google's own backup service for normal backups. When I was rooted I used titanium backup but that was ages ago.

And then a lot of apps that have data that I care about have an export/import functionality, so I can just copy those over without much hassle.

This feature existed as far back as 2014 on my LG G3, so why is it missing on the Pixel?

While there are alternatives on the Play Store, most are either closed-source or bundled into custom launchers — neither of which I wanted. Sending the lock-screen signal requires the Accessibility Service, and I wasn’t comfortable granting that level of access to a closed-source third-party app.

So, I decided to create my own solution. My app is completely open-source, contains no tracking or analytics libraries, and doesn’t make any external requests. The Accessibility Service permission is used solely to send the lock-screen command. I tried to find an alternative to locking the screen without requiring Accessibility permissions but the options either required root or weren't what I needed.

I also saw this as a chance to explore modern Android development. The last time I built an app from scratch, Jetpack Compose didn’t even exist. While I didn’t hate the experience, I won’t say I miss the days of wrestling with XML either. That said, some aspects of Compose felt harder to fully grasp and work with.

You can download it here: https://play.google.com/store/apps/details?id=com.ah.taplock

Source code is here: https://github.com/modelorona/TapLock

Any suggestions or feedback, please feel free to leave here or on the repo.

Comments URL: https://news.ycombinator.com/item?id=42481382

Points: 3

# Comments: 0

Do you have any plans to expand it?

From my understanding, it's used a lot by moderators. It's likely that they may block me since I do not pass any authentication or identifier info, but I don't think it'll happen soon as my use case is extremely simple (knock wood).

I query it as such: https://www.reddit.com/r/earth/top.json?limit=50 and then parse it on the front.

During my testing, I was blocked only once and that was because I ended up querying the same URL too much, which in normal usage shouldn't happen.

The previous iteration used a library called snoowrap (https://github.com/not-an-aardvark/snoowrap) which is unfortunately archived now.

Thank you for the feedback about the image context!

I want to tell you about Scroll-It (https://scroll-it.xyz), a little side project I've been working on. I first made it back in 2020 to learn about VueJS and Vuetify, and I've just finished giving it a big update. You can see the source code on github (https://github.com/modelorona/scroll-it)

Scroll-It is a simple gallery auto-scroller that lets you look through Reddit image posts. You can either go through them manually in the gallery view, or you can start a slideshow from any picture.

I recently rewrote the whole thing to work with newer versions of Vue and Vuetify.

The new version does pretty much the same stuff as before, but it should work faster and use your screen space more effectively. It does not have a backend and instead queries the Reddit JSON API directly, parses it, and displays the images.

While working on this, I got help from AI tools like Claude, Codium, and ChatGPT. Claude was usually the most helpful, especially for tricky questions. Codium was good for small tasks but struggled with complex queries.

I've got some ideas for what to add next, like using keyboard shortcuts, server-side rendering, and lazy-loading images.

I'd love to know what you think about Scroll-It! Add a comment below, or feel free to submit an issue directly on the github.

Comments URL: https://news.ycombinator.com/item?id=41637809

Points: 2

# Comments: 2