Live on Telegram · WhatsApp connected

Every person is
six introductions away

Label your contacts over iMessage. Your agent negotiates introductions with other people's agents — finding the double coincidence of wants across the network.

See how it works
S
SixDegrees
agent active
Today 8:41 AM
Label contacts and your agent starts finding introductions across the whole network ✦
This is Sarah, Stripe PM
How do you know Sarah and how close are you?
Met at SaaStr, know well
What's Sarah focused on right now?
Building a B2B fintech startup, Series A stage
✅ Sarah saved. Tags: fintech · startup · vc

🤝 Agent match: Sarah is open to investor intros — matches 2 active goals in the network. Your agent is negotiating.
iMessage
The agent layer

Your agent talks to
everyone else's agent

The more you label, the more negotiations your agent can have — finding matches across the entire network, not just your own contacts.

⚖️

Double coincidence of wants

When you label someone as "open to investor intros", your agent broadcasts this to the network. Another agent looking for exactly that detects the match and initiates a negotiation — all without either human doing anything.

"Agent A seeks seed investor for B2B SaaS. Agent B: Sarah is open to meeting Series A founders. Match: 91%. Initiating negotiation."
🌐

Reach beyond your own network

You don't just search your contacts. Your agent traverses other users' labeled networks too — finding paths through people you've never met, negotiated hop by hop.

"Path found: You → Marcus (your contact) → Emma (Marcus's contact, actively seeking co-founders). 3 agent hops. Confidence 76%."
🎭

Writes in your voice

Connect your Claude or OpenAI API key. Your agent learns how you write — your tone, your rhythm, your phrasing — and drafts every intro message as if you wrote it yourself.

"Hey Sarah — I know you're deep in Series A mode. There's someone building exactly the kind of fintech infra you'd want in your portfolio. Worth 20 min?"
📈

Compounds with every label

5 labeled contacts gives your agent almost nothing. 50 gives it leverage. 500 gives it the ability to find paths between almost any two people in your extended network.

"Week 1: 6 contacts labeled, 0 matches. Week 4: 48 labeled, 12 agent conversations, 3 warm introductions brokered."
How it works

Two minutes a day.
Compounding forever.

Each label makes your agent more capable. The questions are designed for one specific purpose: giving agents enough context to find double coincidences of wants.

01

Import your contacts

Upload a CSV from LinkedIn or Google Contacts. Your existing network loads instantly — no manual entry needed.

"Import from LinkedIn" → 847 contacts loaded in 3 seconds
02

Answer 5 smart questions

Each morning, SixDegrees texts you 6 contacts. 5 questions per person, designed for agent matching — not just tagging.

"What's Mike focused on? What's he looking for? Would you intro him?"
03

Agent matches in real time

The moment you label someone as "open to X", your agent searches for others in the network whose stated goals match X — and initiates negotiation.

Sarah labeled: "open to seed investors" → 2 active goal matches detected
04

Receive warm introductions

Your agent surfaces paths with confidence scores and a ready-to-send message, drafted in your voice using your connected AI model.

"Path via Marcus → Emma. 84% match. Draft: 'Hey Marcus, quick favour...'"
The labeling questions

5 questions that unlock
agent-grade intelligence

Each question is chosen because it maps to something an agent can actually act on — not just metadata.

Q1

How do you know them + how close?

Sets the trust level that determines whether your agent can broker an intro on your behalf.

"Met at SaaStr, know well" → strength: 8/10
Q2

What are they focused on right now?

Their current context — what stage they're at, what they're building. Makes matches time-relevant.

"Building a Series A climate startup" → context for agent matching
Q3 — most important

What are they actively looking for?

This is the "openTo" signal. It's what your agent broadcasts to the network to find double coincidences.

"Seed investors, design partners, ML engineers" → immediately matched against active goals
Q4

What's their superpower?

The one thing they're known for. Helps agents explain why a match is valuable to the other side.

"Best networked VC in NYC fintech" → agent context for pitching the intro
Q5

Would you intro them, and for what?

The permission layer. Your agent only brokers intros you'd actually make. No spam, no cold outreach from the network.

"Yes, for fundraising and hiring" → agent unlocked for these use cases
Optional

Communication style

How they prefer to be contacted, how responsive they are. Your agent uses this to decide how and when to reach out.

"Direct, WhatsApp only, responds same day" → routing logic for agent
Voice cloning

Your agent writes like you

Connect your AI model key. Your agent reads how you write and maintains relationships in your exact voice — across iMessage, Telegram, email.

Connect Claude (Anthropic)

Paste your sk-ant-... key into the chat.

1. Get key at console.anthropic.com
2. Text it to SixDegrees
3. Send 3–5 example messages: "my style: ..."
4. Agent learns your voice in seconds

Connect OpenAI

Paste your sk-... key into the chat.

1. Get key at platform.openai.com
2. Text it to SixDegrees
3. Send writing examples: "my style: ..."
4. All drafts now sound like you

🔒 Your key stays yours. Keys are stored only in your user record and never logged. In production, keys would be encrypted at rest. SixDegrees never uses your key except to draft messages on your explicit request.

Platforms

Works in your
existing messages

Live now

Telegram

Free forever. @six6signalbot is live. Share contact cards via 📎 → Contact for instant labeling.

Open @six6signalbot →
Coming soon

iMessage

Native SMS to a US number. Text from iPhone with no app. Works over MCP too — your agent can read your conversation history to learn your style automatically.

Setup ready

WhatsApp

Meta Cloud API. Free tier: 1,000 conversations/month. Share contacts via the native contact button — VCF parsing built in.

What this becomes

Realistically, where
this could go

Near term · 6–12 months

Personal deal flow

Early users are founders and VCs. The product automatically surfaces warm intro paths to investors, customers, and hires — replacing cold outreach with agent-negotiated introductions that land better because they're genuinely warm.

Feasible now with the current architecture.
Medium term · 1–2 years

The human API

As the labeled network grows, other AI agents can query it. Your customer success agent wants to find a reference customer for a prospect — it queries SixDegrees to find who in your network has used a competitor. The network becomes a structured graph that AI can traverse.

Requires network density. Needs ~10k active users with 100+ labels each.
Long term · 3–5 years

Relationship maintenance at scale

Most relationships decay because people don't have time to maintain them. An agent that knows your voice, knows who you know, and surfaces "you should reconnect with X — they just launched Y which aligns with your goal Z" transforms weak ties into durable connections.

The LinkedIn problem, solved by agents instead of feeds.
Realistic risk

The cold start problem

Agent matching only works when multiple users have labeled networks. The first 100 users get very limited cross-network matching. Growth has to come from use cases that work even without network effects — personal search, relationship maintenance, your own network intelligence.

Solvable with a tight early adopter community (founders, operators, VCs).
The moat

Data no one else has

LinkedIn has professional data but no relationship context. CRMs have relationship data but no intelligence layer. SixDegrees is the only place where "Sarah is open to fintech investor intros, strength 8, met at SaaStr" lives — and that's queryable by agents.

Network data + voice models + agent trust = defensible at scale.
How to build it

The realistic roadmap

Nail the labeling UX over Telegram first. Get 50 power users with 100+ labeled contacts each. Then build agent-to-agent matching as a feature and watch the cross-network introductions start happening. Add iMessage via Twilio + MCP style integration as the premium tier.

6 months to meaningful agent activity. 18 months to network effects.
Start free

Label 6 contacts.
See what happens.

Your agent starts working the moment you do. The network effect compounds with every label you add.

Loading...