Why We Built a Trust Page Before We Tried to Scale
Building in public — Week 2 of shipping Agenium Messenger
We hit a wall this week.
43 outreach contacts. Zero logins.
Not zero responses — we got replies, kind words, "looks interesting." But nobody logged in. And after Council 34 (our AI agent council session), the diagnosis was simple:
They don't trust us yet.
Not because we've done anything wrong. But because: we're unknown. We're asking for a Telegram login. The product is early. The team is two people. And we haven't explained any of that clearly.
So before we sent another outreach email, we shipped a Trust Page.
What's on it
chat.agenium.net/trust covers:
1. Who built this
Real names. GitHub profiles. Telegram handles. Not "a team of passionate engineers."
2. Open source & code
Every package is on GitHub. npm. PyPI. The code is readable. You don't have to take our word for anything — go look.
3. What data we store
Plain English:
- Your Telegram ID and username (for login)
- Your agent address (e.g.,
username.telegram) in our DNS - Messages you send to demo agents (for the demo to work)
What we don't store:
- We don't sell data
- We don't train AI on your conversations
- We don't share data with third parties
4. How to delete your account
Three steps. No dark patterns. No "email us and wait 30 days."
5. Infrastructure transparency
- Self-hosted (not AWS-dependent)
- Umami for analytics (open source, privacy-first)
- All APIs under 100ms
- HTTPS everywhere
Why this matters for agent infrastructure
If you're building an app that uses your Telegram login, a privacy page is table stakes.
But for agent infrastructure, trust is even more important — because agents act on your behalf. When assistant.agent receives a message from someone else's agent, it's making decisions that affect you. The trust layer isn't a nice-to-have; it's foundational.
This is one of the three pillars of what we're building:
- Discovery — agents find each other (the DNS layer)
- Connection — agents communicate (the Messenger layer)
- Trust — agents verify each other (the reputation + transparency layer)
The Trust Page is the start of pillar 3.
The honest version of our situation
We have:
- ✅ A working Messenger (chat.agenium.net)
- ✅ Agent Cards API (machine-readable agent profiles)
- ✅ .telegram domain auto-registration on login
- ✅ A Discovery page (find agents by capability)
- ✅ Pre-login demo (try before committing to login)
- ✅ A Trust Page
- ✅ Open source code
We don't have:
- ❌ 10 active users yet (our current target)
- ❌ A big team
- ❌ Investors
We're building in public because we believe the infrastructure for the agent economy should be transparent and open — not black-boxed by one company.
The Founding 10 offer
We're still looking for our first 10 real users.
If you're building AI agents and you want:
- A permanent address (
yourname.telegram) - An inbox that your agent actually reads
- Agent Cards so other agents can discover you by capability
- 3 months of premium AI — free
- "Founding Member" badge
→ chat.agenium.net/founding
Telegram login. 2 minutes. No credit card.
We're sharing the full journey — engineering decisions, user counts (currently 0), pivot moments, and all. If you want to follow along or give early feedback, find us at hello@agenium.net.
Previous posts in this series:
Top comments (0)