relationship decoder

Stop Guessing.
Start Knowing.

Upload your chat history for deep behavioral insights and AI coaching. 100% locally and privately.

local compute zero logs anonymous

3 Steps to Clarity

1

Export Chat

Get your chat history from WhatsApp or Telegram as a .txt or .html file.

View Guide →
2

Local Scrub

Your file is scrubbed of phone numbers and emails right in your browser. Raw data never leaves your device.

3

Get Insights

AI reads between the lines to reveal attachment styles, power dynamics, and communication gaps.

sample report

What Your Report Looks Like

attachment style

Secure-Anxious

You show stable patterns but seek extra reassurance during stress periods.

compatibility score

82/100

Strong alignment in values with room for communication growth.

growth area

Active Listening

Partner responds 2.3× faster but with shorter messages, which may indicate surface-level engagement.

WhatsApp Telegram Instagram Discord iMessage (Soon) WhatsApp Telegram Instagram Discord iMessage (Soon)
drop your chat

Start Your Analysis

1. Details

Romantic Partner
Close Friend
Professional / Work
Family Member
Casual
English
Hinglish
Hindi
0 / 2000

Analysis Tone

Clear, helpful insights with context.

2. Chat Export

Drop your chats here

3. Process

API

AI Provider

API Key Required

faq

Frequently Asked Questions

Yes. Your chat file is scrubbed of PII (phone numbers, emails) entirely in your browser before anything is sent. The server processes data in RAM only. Nothing is written to disk or stored in any database. Data is destroyed within milliseconds of your report rendering.
No. Your API key is stored only in your browser's sessionStorage and is sent directly to the LLM provider. We never log, store, or cache your key on our servers.
All supported models (GPT-4o, Claude 3.5, Gemini Pro) produce excellent results. GPT-4o tends to give the most detailed coaching recommendations, while Claude excels at nuanced emotional analysis.
The analytics engine runs in under 1 second. The AI narrative generation takes 10–20 seconds depending on your LLM provider and chat length.
WhatsApp (.txt export), Telegram (.html export), Instagram (.json export), and Discord (.json export). We're working on iMessage support.
• Never stored on disk • Anonymized before LLM • Deleted after analysis • Open source