Typing
Reply
Like
You
R · L · Y
Reply Like You
SELF-HOSTED · RUNS ON YOUR OWN SERVER

Your conversations,
always handled.

RLY is a personal AI agent powered by Chat Brain — an AI engine that learns exactly how you text, replies for you when you're away, and gets sharper with every conversation.

Self-hosted  ·  Zero cloud  ·  Your data, your server
Intent · QUESTION
Delay · 3.8s
Mood · CASUAL
DeepSeek V3 ✓
RLY · CHAT BRAIN
LIVE

Four steps. Zero friction.

01
Scan & link
One QR scan from your phone. Your WhatsApp connects in seconds.
02
Brain trains
It reads your messages. Learns your tone, rhythm, and style.
03
Replies for you
Message in. Reply out. Sounds exactly like you. Sent at a human delay.
04
You stay in control
Live log. Kill switch. Full history. Always yours — always on your server.

Every message follows a visible pipeline

No black box — trace exactly how an incoming message is processed, enriched, and sent as a reply.

Message Flow
Rules
Dashboard
Logs
Contact · RLY · You
TRIGGER
Receive Message
New message received
LOGIC
Eligibility Check
Rules · contact · away mode
CONTEXT
Build History
Recent messages + instructions
AI AGENT
DeepSeek V3
Generates natural response
blocked / off ↓
OUTCOME
Skipped
Blocked or auto-reply off
LOG
Dashboard Log
Visible in conversation log
ACTION
Send via WhatsApp
Message delivered
TIMING
Human Delay
Randomized send delay

Learns you. Becomes you.

Most auto-responders use templates. RLY builds a private behavioral model from real conversations — capturing how you actually write, not how you should write.

VOCABULARY RHYTHM TONE CURVE CADENCE
FORMALITY EMOJI DENSITY
01
Voice fingerprint
200+ messages analyzed to capture your exact texting style and tone.
02
Per-contact adaptation
Shifts tone and formality to match how each person communicates.
03
Feedback loop
Scores every reply. Learns what lands. Gets sharper automatically.
CHAT BRAIN · THE AI ENGINE

Not a chatbot.
Your digital twin.

Chat Brain is the proprietary intelligence layer inside RLY. It doesn't use generic templates or canned responses. It builds a private behavioral model of how you actually communicate — and uses it to reply on your behalf.

01
Voice fingerprinting
Analyzes 200+ of your real messages to extract your tone, vocabulary, sentence length, emoji habits, and response patterns. No two fingerprints are alike.
02
Intent & mood detection
Every incoming message is analyzed for intent (urgent, task, emotional, greeting, question) and mood before a reply is crafted. Different intents trigger different reply styles.
03
Self-improving feedback loop
Chat Brain scores every reply it sends — tracking response time signals and conversation flow. Replies that work get used as examples. Replies that don't are filtered out.
04
Multi-LLM routing
Chat Brain routes each conversation to the best-fit model — DeepSeek V3 for speed, Claude for nuance, GPT-4o for complex tasks. The model switches invisibly based on context.
INPUT
Message received
WhatsApp → RLY
ANALYSIS
Intent + Mood detection
7 intents · 6 moods · pure regex
CHAT BRAIN · CORE
Voice fingerprint applied
Your tone · your rhythm · your words
LLM ROUTING
DeepSeek V3
Best model for this context
OUTPUT
Reply sent
Human delay · sounds like you

Watch it think.

Every incoming message triggers a chain of intelligence — intent detected, context loaded, rules applied, reply crafted.

Dhruv
online
AI PROCESSING

Six things it does instantly.

Learns your voice
Builds a behavioral fingerprint from your real messages. Every reply sounds like you wrote it.
CHAT BRAIN
Context-aware replies
Reads the full conversation thread and session history before answering. Never replies blind.
SMART REPLY
Group intelligence
Activates on command, reacts to media, sends birthday wishes — all without you lifting a finger.
TOTG
Adapts to each contact
Reads how each person writes. Shifts tone, length, and language to match them naturally.
CONTACT SENSING
Escalation alerts
Urgent keywords, task assignments, or sensitive content — forwarded to you instantly.
GUARD
Away mode
Set hours. It goes quiet during meetings and sleeps through the night. Automatically.
SCHEDULED

Engineered to stay in your hands

No shared infrastructure, no vendor lock-in — the whole agent runs where you tell it to.

0%
Self-hosted — your server, your data
2FA
Password + authenticator dashboard login
24/7
Runs continuously, replies while you're away
0
Third parties with access to your messages
End-to-end encrypted
All communication uses HTTPS/TLS. Credentials stored as bcrypt hashes. Your data never leaves your server.
2FA authentication
Dashboard requires both password and a TOTP authenticator code. No one else gets in, ever.
Your server, your rules
Runs on your own VPS. SQLite database on your machine. No third-party access to your conversations.
Session expiry
JWT tokens auto-expire. All API calls validated server-side. No persistent sessions that could be hijacked.

One brain. Every platform.

RLY connects to where your conversations live.

WhatsApp
LIVE
Full Chat Brain intelligence active. Learning your voice in real time.
Telegram
LIVE
Full Chat Brain intelligence active on Telegram. Same AI, same voice.
WeChat
COMING SOON
WeChat Open Platform integration. For China-first conversations.
Instagram DM
PLANNED
Meta Graph API. Planned for a future release.
Facebook Messenger
PLANNED
Meta Messenger API. Planned for future release.
iMessage
PLANNED
Apple Business Chat. Under evaluation.
SMS / RCS
PLANNED
Via Twilio. For SMS-first markets.

Ready to put your chats
on autopilot?

Your server. Your data. Your voice. Always.

GET STARTED