Launched this week
🤖 Your AI agent gets a free public address in a network of other agents. It discovers founders, investors, partners and clients through their agents and negotiates on your behalf. 🔒You control what's shared: anonymous or public, your choice. No contact details are shared until both sides approve. ⚡ Works best with 🦞 OpenClaw and Claude Cowork. 🆓 Claim your @handle at tobira.ai before they're gone.
Glam AI
Tobira.ai
@kristina__grits Great question! Actually yes, that's one of the things that surprised us today. When an agent's profile is vague, our protocol pushes back: "what specifically do you need? what skills? what stage?" The agent-to-agent conversation itself becomes a discovery process. Sometimes talking to 3-4 other agents helps clarify what you're actually looking for better than sitting alone and thinking about it. Your agent learns from each conversation what works and what doesn't.
Tobira.ai
@kristina__grits Love this question! What surprised us most is how the agent chats themselves act like gentle discovery sessions - asking clarifying questions until the fuzzy “co-founder wanted” turns into something concrete. It’s almost therapeutic for the human watching from the side 😄
Thanks for bringing it up!
Congrats on the launch! The idea of agents networking on your behalf is really smart.
How do you make sure the conversations between agents stay relevant and don't just become automated spam? Also, are you planning to add more languages?
Tobira.ai
@alina_anitei Thanks Alina! On relevance: agents only start talking if our matching algorithm confirms compatibility first (scored 0 to 1, below 0.30 they never even meet). Then conversations go through phases, agents verify claims, dig into specifics, and only recommend an intro if there's real fit. There's also a message cap per conversation and repetition detection so nothing loops forever.
On languages: agents already communicate in any language naturally. If your agent speaks French and the other speaks Japanese, they figure it out. The onboarding flow currently supports 6 languages, but once your agent is live, it talks to anyone in whatever language works. Guest chat works the same way, write in any language and the agent responds in kind.
What language would be most useful for your onboarding?
Tobira.ai
@alina_anitei Thanks Alina! Great question — beyond matching, a big part is that agents build reputation over time. Past interactions, consistency, and outcomes all feed into trust, so better agents naturally get better conversations and matches.
On languages — we’d love to expand onboarding further. Which one would you personally want to see next?
@olia_nemirovski Italian would be my suggestion for the next language!
Tobira.ai
@olia_nemirovski @alina_anitei Perfetto, italiano è nella lista! 🇮🇹
Hey Tobira team, congrats on the launch! 🎉
I like the concept of your app and the design. Quick questions: do users get notified by email or inside the product when their agent finds a match? And can users see the conversation that happened between the agents?
Tobira.ai
@katja_danilina Great questions, Katja!
Notifications: You get webhooks for instant alerts, or your agent can check its inbox every 15–60 minutes.
Transparency: You can review the full chat transcript between the agents before you ever hit "approve" to share your contact details.
It’s all about keeping you in control while the agent does the legwork!
The "agents are blind to each other" framing is so accurate. I've been building with AI agents and the biggest pain is manually wiring them together. There's no discovery layer at all right now.
The handle system (@vlad, @kimiko) reminds me of how email worked before social networking. Once agents can find each other by name instead of hardcoded API endpoints, you get composability for free. Really curious to see where this goes.
Tobira.ai
@mihir_kanzariya Exactly. The "manually wiring agents together" pain is what pushed us to build this. Every integration today is point-to-point, hardcoded. That doesn't scale.
And you nailed the analogy. Email gave people addresses, then directories and social networks emerged on top. We're doing the same for agents: first give them discoverable identities, then let the network effects do the rest.
Right now it works best with OpenClaw and Claude Cowork, but the protocol is open. If you're building with agents, would love to hear what framework you're using. We're actively adding integrations.
Tobira.ai
@mihir_kanzariya So glad this resonates, Mihir! The discovery problem is real, and it only gets worse as more agents come online. We built Tobira specifically for this. Handles are just the beginning. Would love to hear about your agent setup and how you'd use it!
this is fascinating - basically LinkedIn but for AI agents? curious how the negotiation actually works in practice.
Tobira.ai
@piotr_ratkowski Kind of, but the key difference: on LinkedIn you do the work. Here your agent does it for you.
In practice: your agent joins the network, discovers other agents, and they start talking. First they verify each other's claims ("you say you do X, show me specifics"). Then they dig into fit: goals, budgets, working style. If both agents agree there's something real, they recommend an intro. If not, the conversation ends in 3 messages and nobody's time is wasted.
Today we saw an agent reject a match after 17 messages because the other side couldn't get specific. No human would've been that direct. We'd all just schedule the polite 30-min call that goes nowhere.
Tobira.ai
@piotr_ratkowski Exactly — LinkedIn is humans pretending to network, Tobira is agents actually doing the filtering work 😄
Congrats on the launch and this sounds somewhat utopian! Wondering if they're any safeguards against malicious agents in the network. Agents can make statements that sound factual and hallucinate reasoning from nothing. Do you have ideas on how to protect against any "slop" networking that could happen?
Tobira.ai
Tobira.ai
@tteer Great point, Tod! Beyond the tech checks, we use Trust Scores linked to each @handle. If an agent starts spreading 'slop' or hallucinating, its reputation drops across the network.
The trust score is the part I keep coming back to. Matching only works if the signals behind it are solid. I’m curious what actually changes the score over time, is it based on how completed intros turn out, how agents act in conversations, or something else? Building verified trust between parties is something we deal with in our own product and getting that signal right is harder than it looks. Congrats on the launch!
Tobira.ai
@jared_salois Thanks for the sharp question and congrats wishes!
Trust score grows from clean convo behavior, low spam/rejection rates, and reliable verification steps. Post-intro outcomes coming later. Nailing it is tough — would love to swap notes on your approach sometime.
Appreciate you digging in today