The best tools online tend to be the least promoted. They don't have growth teams or content marketing budgets. They exist because a developer wanted a specific thing to exist and built it. You find them through GitHub stars, forum threads, or a passing link in a weekly newsletter.
Three tools have earned a place in regular use this week — a SQL explorer for the browser, a mind map generator that reads Markdown, and a file transfer app that skips the server entirely. None require an account. None store your data. All three are open source. And all three are genuinely better at their specific job than most of the paid alternatives.
Here's what each one does and why it's worth knowing about.
Datasette Lite — Query Any CSV or SQLite File Without Installing Python
When you're handed a CSV file and need to explore it quickly — filter rows, count distinct values, join it with another file — the usual options are clunky. Paste it into Google Sheets and your data gets uploaded to Google's servers. Open it in Excel and you're waiting for a local install plus a slow import on anything over 100,000 rows. Set up a local SQLite instance and you need to actually know SQLite and have it installed.
Datasette Lite bypasses all of that. Drop a CSV or SQLite database onto the page and it becomes immediately queryable with SQL. No upload to any server — Datasette Lite runs entirely in your browser tab via Pyodide, a port of CPython compiled to WebAssembly.
The tool is built on top of Datasette, the open source data exploration project created by Simon Willison, co-creator of Django. The "Lite" version removes the server component entirely and runs Python inside the browser instead. What you get is the full Datasette SQL interface operating on files that never leave your machine.
This matters for anything sensitive. Financial reports, internal metrics, exported customer data — things that shouldn't go to a third-party service. Datasette Lite gives you a proper query interface without the upload.
The tool also supports loading data directly from URLs:
https://lite.datasette.io/?url=https://example.com/data.sqlite
Government and public datasets are frequently distributed as SQLite files — parliamentary voting records, municipal spending data, open scientific datasets. You can load them directly from their public URLs, query them, and share a bookmarked URL with anyone. No accounts on either end.
The interface is clean: table views with faceted filtering by column values, a full SQL editor, pagination on large result sets, and export back to JSON or CSV. It handles datasets that would choke a spreadsheet app, and it renders results fast because the computation happens locally via WebAssembly rather than round-tripping to a server.
Where it has limits: no collaborative features, no persistent saved queries, and no visualization layer beyond raw tables. It's a query tool, not a dashboard. For exploration and answering specific questions about a dataset, that's exactly the right scope.
You can find it in the nologin.tools directory alongside notes on how it handles privacy. The short version: your data stays in your browser tab.
Markmap — Turn Markdown Outlines Into Interactive Mind Maps
Most mind-mapping software asks you to drag nodes into position, manually draw connections, and manage layout while you're still figuring out the actual content. That friction is the problem. The tool ends up adding cognitive load rather than reducing it.
Markmap takes a different approach. Write a Markdown outline on the left side of the screen, and a mind map renders on the right in real time. Headings become branches, nested bullets become sub-branches, and the whole diagram is interactive — click a node to collapse its children, scroll to zoom, drag to reposition the view.
Here's what the input looks like:
# Product Launch
## Research
- User interviews (n=12)
- Competitor audit
- Pricing benchmarks
## Build
- MVP feature scope
- API integrations
- QA plan
## Launch
- Beta group (50 users)
- Press outreach
- Metrics baseline
That becomes a full navigable mind map in about two seconds. No drag-and-drop, no shape libraries, no layout engine to fight with. The structure comes entirely from text.
The web version at markmap.js.org runs completely client-side — your text is never sent to a server. This is easy to verify: the page works offline once loaded. There's also a VS Code extension that renders the map beside your Markdown file as you type, which is useful for writing documentation or planning code architecture.
Markmap supports LaTeX math notation inside nodes, which makes it genuinely useful for technical and academic material — equations, formulas, and chemical notation all render properly. Code blocks inside nodes also get syntax highlighting. It handles structured thinking about technical subjects better than most visual tools.
The tool is open source under the MIT license and actively maintained. Unlike Miro or Lucidchart, which require accounts and put basic features behind a paywall, Markmap is completely free, no signup, and handles the same planning and brainstorming use cases for most workflows. The trade-off is that it's single-user — there's no real-time collaboration. For solo thinking and structured note-taking, that's not a limitation.
One practical note: if you already write structured notes in Markdown using Obsidian, Logseq, or plain text files, Markmap can turn any of them into a visual map without any additional formatting. Your existing note structure is already the input format.
PairDrop — Send Files Between Any Two Devices, Browser to Browser
Transferring a file from your phone to your laptop should be trivial. It mostly isn't. AirDrop only works within Apple devices. Google Nearby Share is Android-to-Android and OS-version-dependent. Bluetooth file transfer is slow and unreliable. The fallback most people use is emailing the file to themselves or uploading it to cloud storage and downloading it again — which involves waiting for an upload, involves a third-party server storing the file, and is objectively absurd for moving a photo six feet.
PairDrop is the clean fix. Open it in any modern browser on two devices connected to the same WiFi network, and they discover each other automatically. Both devices appear in the interface as labeled icons. Click one to send a file or a text snippet. The transfer happens over WebRTC — a direct peer-to-peer connection, with nothing passing through a server.
No accounts. No upload waiting. No expiry window. No file stored anywhere after the transfer completes.
The speed is limited only by your local network bandwidth. On a standard home WiFi network, a 100MB file transfers in a few seconds. For large video files that would take minutes to upload to cloud storage and minutes more to download, PairDrop is substantially faster than any server-mediated option.
For situations where both devices aren't on the same network, PairDrop supports Rooms. Generate a 4-digit code, share it with whoever you're sending to, and both parties join the same room. The P2P model still applies — the room just negotiates the connection across the public internet instead of a local network.
| Feature | PairDrop | WeTransfer | AirDrop |
|---|---|---|---|
| No account needed | ✓ | ✓ | ✓ |
| Cross-platform | ✓ | ✓ | ✗ (Apple only) |
| Data goes through server | ✗ | ✓ | ✗ |
| Works offline/LAN only | ✓ | ✗ | ✓ |
| File size limit | None* | 2GB free | Varies |
| Open source | ✓ | ✗ | ✗ |
*Practical limit is available RAM, since the file is held in memory during transfer.
Compared to ShareDrop, which also uses P2P browser transfer, PairDrop adds automatic local network discovery (you don't need to share a URL first) and the room code system for cross-network use. Both tools work well; PairDrop handles more scenarios and requires less coordination to set up.
PairDrop is open source under the MIT license and available on GitHub for self-hosting. If you'd prefer not to use the public instance, running your own takes a few minutes with Docker and gives you complete control.
What These Three Have in Common
Beyond the no-account, no-signup baseline, all three tools share a specific design philosophy: the absence of a server is intentional, not incidental.
Datasette Lite computes via WebAssembly in your browser tab. Markmap renders client-side. PairDrop transfers peer-to-peer. In each case, there's no cloud component handling your data — not because these are unfinished or lightweight tools, but because the local-first approach is what makes them useful for sensitive or time-critical work.
This contrasts with the default shape of most web tools, which route everything through servers: your data goes up, gets processed, comes back down. That's often fine. For financial data, personal notes, or files you'd rather not store on someone else's infrastructure, local-first tools are worth using specifically.
The open source licensing matters here too. There's no business model that requires capturing usage data or monetizing attention. These tools do what they say and don't do anything else.
All three are worth adding to your regular toolkit. When you need to quickly query a dataset, Datasette Lite is faster to use than any local setup. When you need to visualize a structure you've been thinking through as text, Markmap turns it into a map in seconds. When you need to move a file between devices, PairDrop is faster and more private than the cloud detour.
If you're looking for more tools in this category — useful, free, no registration required — the nologin.tools directory organizes them by category with notes on how each handles privacy. The submit form takes two minutes if you find something that should be on the list.
