Launching today
Tictable
The spreadsheet rebuilt for AI
26 followers
The spreadsheet rebuilt for AI
26 followers
Tictable is the best place to work with small data, big data and everything in-between. Tictable runs SQL directly in the browser and lets AI explore your data by executing hundreds of queries in real time. Ask questions, generate live reports, and turn messy data into insights, all while keeping the simplicity of a spreadsheet.
Tictable
Hey Product Hunt 👋
We’re two brothers from Sweden building Tictable.
We built Tictable because spreadsheets are still the best way to work with data. But they start to break when you want deeper analysis, better performance, or AI that actually understands what you are doing.
Tictable is a fast, minimal spreadsheet with a real query engine. You can use SQL or just ask with AI, directly in the browser.
A few things we think make it special:
• SQL runs directly against your spreadsheet data
• AI can generate filters, formulas, and reports from plain language
• Charts and reports stay in sync as your data changes
• File columns keep attachments tied to each row
• It stays fast, even with large datasets
Recent updates include live reports, smarter imports, custom SQL queries, formula columns, improved filters, a sidebar agent, faster rendering, file support, kanban boards, stronger charts, and split views.
We’d love feedback on a few things:
1. What would make you switch from Airtable, Sheets, or Notion?
2. Is it more important to analyze existing data or manage your own inside the tool?
3. What should we focus on next?
Happy to answer anything all day 🙌
I often work with CSV, Excel, and JSON.
Can ticktable Import all these formats directly in the browser? Handle large files without crashing? Detect and fix common data issues (missing values, inconsistent dates) on import? Connect to cloud storage for direct import?
Tictable
@elena_nimchenko Hey! We got Excel and CSV covered today. I need a few more days to get JSON (and Parquet and Arrow) support ready. Importing via URL is part of that PR as well. I can have that merged into the product this week if you're interested! 🏃♀️
Yes, Tictable handles gigantic files without crashing. We have an actual SQL database running inside your browser and it scales very nicely. When you import data, the flow is like this:
- Tictable puts it in a temporary table that the agent has access to
- You can talk to the agent to massage the data ("drop column x", "normalize column y"...). The agent runs actual SQL on your temporary table so that it can update millions of rows insanely fast without using AI tokens.
- When you're done, click "save" and the massaged table gets saved.
Hope that helps! Happy to answer any other questions or jump on a call/WhatsApp if you wanna go deep on this!
PS. Here's an example of massaging 1 million rows with the agent:
If we are running the SQL in browser, How does it affects the performance? I mean I have worked with files of millions of rows and hundreds of columns can it handle that?
Tictable
@nayan_surya98 It should be all good! We use DuckDB under the hood. It uses WebAssembly and runs in a separate worker to ensure that the UI never freezes up. Let me know if you want some benchmarks :)
Features.Vote
most spreadsheet AI tools sit alongside the data and suggest what to type next, which means you're still the one executing, just with better autocomplete.
running SQL directly in the browser and letting the AI fire hundreds of queries in real time flips that completely. the AI isn't advising, it's actually doing the exploration. the gap between asking a question and getting the answer stops being "write a formula and see if it works" and becomes just... the answer.
Tictable
@gabrielpineda Agreed! 😀 This has been great for us — token usage remains minimal and because agents know SQL well out of the box, they produce very good and interesting queries in 99% of the cases.
A tangent, but for our "Live Report" feature, we do exactly this:
- The user enters a prompt like "Tell me how my EMEA sales are progressing. Give me a report in Swedish."
- An agent explores the data and generates a bunch of interesting SQL queries matching the user's prompt.
- We store those queries in the DB.
- When you go to the "Live Report", we execute all those queries in <1 second and pass the result to a cheap and fast model that converts the results to a readable format.
It works really well!