Tabwise

AI data analyst that saves you 20+ hours a week. Turn messy spreadsheets into structured reports instantly. Talk to your data, cut 90% of the grunt work.

[Abhishek Hegde @abhegd]

🚀 Just shipped for preview https://t.co/pmm1Mlmkzm - AI data analyst.
Turn messy spreadsheets into leadership-ready reports instantly. No formulas, no SQL - just ask questions in plain English. Talk to your data, cut 90% of the grunt work.

Here is a demo with real retail data:
📈 Weekly sales trends across 3 years of store performance
🏪 Store performance rankings with consistency analysis
🎄 Holiday impact analysis with statistical significance testing
📊 Economic correlation analysis (unemployment + CPI effects on sales)

Hours of manual analysis work done in minutes, with structured insights, charts, recommendations, and data you can immediately share with stakeholders or leadership

@tabwiseai is free to try, Give it a spin, burn my token credits → https://t.co/pmm1Mlmkzm

[Abhishek Hegde @abhegd]

Trying @capcutapp for the first time, demo vid incoming, let's go 🚀

[Abhishek Hegde @abhegd]

Shipped @tabwiseai Your AI Data Analyst. Think Cursor for Data Analysis and BI. Add your data, ask for analysis, and get detailed results with charts.

Here is the demo → https://t.co/fXFOCu3Npp
Give it a spin. https://t.co/Jkn6PV4s8U

[Abhishek Hegde @abhegd]

Trying Fast vs Deep analysis in @tabwiseai, giving users more control over analysis mode,

Fast mode runs light models for quick insights.
Deep mode brings in the heavy hitters (Claude, GPT-5 soon?) for rich, structured analysis.

Not every question needs a rocket engine, but some do!
Ready for GPT-5? Let’s bring it 🚀

[Abhishek Hegde @abhegd]

In July alone, I edited 110,000+ lines of code with AI.
Got curious and pulled my own @cursor_ai usage CSV.

Took it to the AI data analyst I built and asked a few questions about my coding habits.

Here’s what I found:
– Heavy sessions = 59.2% of costs, but only 25% of sessions
– Top 25% of sessions use 74.9% of all tokens
– Heavy sessions cluster on specific days → intense work spikes
– Peak usage: Monday ($12.77 avg)
– Lowest: Sunday ($2.95 avg)

You should be able to talk to your data.

[Abhishek Hegde @abhegd]

I got the new @OpenAI oss models running for Tabwise this morning. gpt-oss handles simple analysis well but stays surface-level for now.
Testing, Qwen3 Coder 480B, DeepSeek V3, Llama 4 Maverick Instruct as well for simple tasks. Powered by
@FireworksAI_HQ

Here I picked the model, but ideally, you shouldn’t have to! Tabwise auto-routes based on task depth and latency

[Abhishek Hegde @abhegd]

I got the new @OpenAI oss models running for Tabwise this morning. gpt-oss handles simple analysis well but stays surface-level for now.
Testing, Qwen3 Coder 480B, DeepSeek V3, Llama 4 Maverick Instruct as well for simple tasks. Powered by @FireworksAI_HQ
Here I picked the model, but ideally, you shouldn’t have to! Tabwise auto-routes based on task depth and latency

[Abhishek Hegde @abhegd]

Improved context management. The analyst now really understands the data, responses are sharper, more accurate. Probably over engineering it at this point.
It’s been stuck on localhost for too long...

[Abhishek Hegde @abhegd]

You can ask a follow up question, after your analysis