DeepTable – an API that converts messy Excel files into structured data
We tried to build an Excel error checker. To achieve that, we needed to actually understand the semantic structure of a spreadsheet first. So we built that, and it turned out to be the harder, more general problem. The core issue: most real-world spreadsheets aren't relational tables. Merged cells, multi-level headers, multiple tables per sheet, totals mixed in with data. You can't just dump them to CSV and call it done. LLMs handle the easy cases but fall apart on complex workbooks at scale. Our approach uses an agent-guided compilation pipeline that produces SQL-ready relational tables…
What it does
In the maker’s words, at launch
We tried to build an Excel error checker. To achieve that, we needed to actually understand the semantic structure of a spreadsheet first. So we built that, and it turned out to be the harder, more general problem. The core issue: most real-world spreadsheets aren't relational tables. Merged cells, multi-level headers, multiple tables per sheet, totals mixed in with data. You can't just dump them to CSV and call it done. LLMs handle the easy cases but fall apart on complex workbooks at scale. Our approach uses an agent-guided compilation pipeline that produces SQL-ready relational tables with full cell-level provenance. This demo visualizes what we do: https://storage.googleapis.com/deeptable-public/deeptable_an... We have a handful of early customers but honestly don't know yet whether this is a real market or a niche problem. We're posting this to hear from people who've dealt with arbitrary spreadsheet ingestion. Whether you solved it, gave up, or are still living with the pain. If you want to try it on your own files, email me (see my profile for my email) and I'll give you API access.
Does the same job
all alternatives →- IMI made a tool to convert images of tables to CSV2021 · github.com · ▲354
- SASpreadsheetDB – A database that you can query with spreadsheets2017 · spreadsheetdb.io · ▲144
- STSheetlabs – Turn your spreadsheets into APIs2014 · sheetlabs.com · ▲158
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, March 2026
the whole month →
Switch from ChatGPT to Claude with import memory feature
AI · Mar 2026 · claude.com
