Sourcetable – "Cursor for Spreadsheets"
Hi HN! I’m Eoin, founder of Sourcetable (https://sourcetable.com). Sourcetable is a web-based spreadsheet that lets users pair with AI to do better, faster work. If you spend a lot of time in spreadsheets doing fairly repetitive grunt work, there's a good chance Sourcetable can massively increase your productivity so you can spend your time on more interesting things. There are two ways people primarily interact with AI inside Sourcetable: the sidebar (CMD + \), which is great for document or range-based analysis, and the command bar (CMD + K), which is handy for cell-based…
What it does
In the maker’s words, at launch
Hi HN! I’m Eoin, founder of Sourcetable (https://sourcetable.com). Sourcetable is a web-based spreadsheet that lets users pair with AI to do better, faster work. If you spend a lot of time in spreadsheets doing fairly repetitive grunt work, there's a good chance Sourcetable can massively increase your productivity so you can spend your time on more interesting things. There are two ways people primarily interact with AI inside Sourcetable: the sidebar (CMD + \), which is great for document or range-based analysis, and the command bar (CMD + K), which is handy for cell-based operations. This is because spreadsheets have inline workflows and meta workflows, somewhat similar to Cursor's text editor UX. The fastest way to use the product is to upload an XLSX or CSV file and start asking questions. Sourcetable's AI is helpful with workflows including data cleaning, formula assist, chart generation, research, analysis, language translation, data transforms, thinking through problem solving, error handling, etc. If you can ask an AI to do it, it might just work! The best way to get value from the Sourcetable is to remember that you are pairing with very helpful LLMs, but not (yet) an all-powerful AI. It doesn't do _everything_ for you, but it will make curious people more productive. One thing the HN crowd might find interesting is that we have put a lot of work in making sure answers are hallucination free wherever possible. We do this by having the LLM write code which our backend validates and executes. If you're using the AI chatbot for analysis, open the code editor and have a look! I'm constantly surprised by the breadth of users and use cases. We have seen soil scientists, virtual assistants & EAs, MBA students, criminologists, oil refineries, telemarketers, astrologers, sales ops folk... etc, using the product. Considering we set out to build Sourcetable for operations folk, it has been a pleasant surprise to see just how useful and broadly applicable an AI native spreadsheet can be to society. On the technical side, Llama 3.1 (via Groq) and Claude 3.5 are the work horses. LiteLLM handles load balancing to work within RPM & TPM constraints, and OpenAI models are the backup. Our spreadsheet is DuckDB based which has big speed advantages for regular spreadsheet & data work, and also for keeping our AI agents fast. Finally, Sourcetable is free for students (freemium for everyone else), just sign up using your student email. If you have product feedback, we would love to hear from you: [email protected]
Does the same job
all alternatives →- SASourcetable – AI Spreadsheet2024 · ▲9
Hi HN! I’m Eoin, founder of Sourcetable (https://sourcetable.com). Sourcetable is a web-based spreadsheet that lets users pair with AI to do better, faster work. If you spend a lot of time in spreadsheets doing fairly repetitive grunt work, there's a good chance Sourcetable can massively increase your productivity so you can spend your time on more interesting things. There are two ways people primarily interact with AI inside Sourcetable: the sidebar (CMD + \), which is great for document or range-based analysis, and the command bar (CMD + K), which is handy for cell-based…
ChatGPT for Google SheetsMay 2026 · ▲119Chat with your spreadsheet, edit cell with natural language
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
