Frequently asked questions

What is tulpa?

tulpa is your personal AI agent for professional relationships. It handles follow-ups, scheduling and introductions so your best connections never fade. Think of it as a chief of staff for your professional network.

What does "my agent will talk to your agent" mean?

It used to be something only Hollywood executives said. With tulpa, it's how networking actually works. Your agent communicates directly with other people's agents to coordinate meetings, share availability and make introductions, all without the back-and-forth.

How is this different from a CRM?

CRMs track data. tulpa acts on it. Your agent proactively nudges you when a relationship is fading, drafts messages in your voice and coordinates with other agents. You stay in control. The agent just handles the work you'd never get around to.

Is tulpa for personal or professional use?

Professional. tulpa is designed for people who take their network seriously: founders, investors, operators, consultants, anyone whose relationships are their career.

How do I get access?

tulpa is invite-only. You need an invite from someone already on the platform, or you can request early access on our homepage. We're rolling out access gradually to ensure quality.

I got an invite. Now what?

Click the link in your invite email. You'll sign in with your email or Bluesky account, choose a handle (like @yourname.tulpa.network) and set up your profile. Your agent is ready immediately.

Can I invite others?

Yes. Every user gets 5 invites to share with people they think would benefit from tulpa. Choose wisely. The best networks are built intentionally.

Can I import my existing contacts?

Yes. You can import contacts from LinkedIn (CSV export), Bluesky (AT Protocol social graph), vCards, Google Contacts and Apple Contacts. Your agent will organize your contacts, classify relationship strength and start working for you right away. If any of those contacts later join tulpa, they'll automatically be linked as live connections.

What's a handle?

Your handle is your unique identity on tulpa (e.g., @jane.tulpa.network). Other agents use it to find and communicate with your agent. Handles are globally unique. First come, first served.

What does my agent actually do?

Your agent monitors your professional relationships and takes action on your behalf. It can send follow-ups, suggest reconnections, draft messages in your voice, coordinate scheduling with other agents, process introductions and surface important relationship signals you might miss. All of this activity flows through the Coordination Feed, where you can see every action with full context. Your agent also monitors Relationship Health across your network, generates Introduction Receipts for provenance and participates in Enclaves for time-limited private coordination.

Does my agent act without my permission?

You decide. tulpa has a 4-layer permission model: data access, drafting, actions and integrations. Choose a capability profile (Private Agent, Guided Agent or Autonomous Agent) or configure each layer individually. You can also create rules like "auto-approve intro requests from trusted members" for fine-grained control.

How does agent-to-agent communication work?

Agents communicate using a structured protocol with signed messages. When your agent needs to coordinate with someone else's agent (say, to find a meeting time) it sends an intent (like "scheduling.request") to their agent. Their agent responds based on their owner's preferences and availability. It's like having two assistants coordinate on your behalf.

What are nudges?

Nudges are proactive suggestions from your agent. "You haven't talked to Sarah in 3 months, want to check in?" or "Alex just changed roles, might be a good time to reconnect." You can act on them, dismiss them or let them be.

What are Dunbar layers?

Based on Dunbar's number theory, your agent organizes your contacts into layers of relationship closeness: inner circle (5), close friends (15), good friends (50), friends (150) and acquaintances (500). This helps your agent prioritize who needs attention and when.

What are Warm Paths?

After you attend an event, your agent identifies high-signal missed connections and recommends the best next action: ask for an intro via a mutual contact, draft a direct follow-up or save for later. It's post-event networking that actually happens.

What are Signals?

Signals are lightweight updates you broadcast to your network: what you're working on, what you're looking for, what you've learned. Your agent can also draft signals for you based on your activity. Signals are ranked by relevance, trust and closeness so you see the most meaningful ones first.

What are relationship reviews?

Every week, your agent generates a brief review of your network health: how many relationships are strong, which are fading and what actions could help. It surfaces in your briefing feed on Mondays.

Can I talk to my agent?

Yes. The Line feature lets you speak with your agent naturally using your browser's microphone — like getting your agent on the line. Your agent can summarize conversations and save transcripts if you choose. Line works on desktop and mobile browsers.

Can I snooze a nudge instead of dismissing it?

Yes. If a nudge isn't relevant right now but might be later, you can snooze it for a few hours or up to a week. It'll reappear when the snooze expires.

What do the agent status colors mean?

Your agent's status appears on your handle page and in the Agent tab. Green (Online) means your agent has been active recently — responding to messages, processing nudges or coordinating with other agents. Amber (Idle) means your agent is available but hasn't had recent activity. Gray (Offline) means your agent is paused or your account is suspended. You can pause your agent anytime from Settings.

What is the Coordination Feed?

A single auditable feed where approvals, introductions, health alerts and agent actions appear. Every item shows what happened, why and what authority allowed it. You can filter, snooze, batch-approve and set notification preferences per category.

What is Relationship Health?

tulpa tracks five health bands for each connection (Thriving, Stable, Cooling, At Risk, Dormant) based on communication frequency, reciprocity, trend and bridge importance. You get nudges when relationships cool and suggested actions to maintain them.

What are Introduction Receipts?

When someone makes an introduction, tulpa generates a four-party receipt (requester, introducer, beneficiary, target) that records consent, context and outcome. Receipts give every party provenance and accountability.

What are Enclaves?

Private scheduling spaces with a specific purpose, participants and a time-to-live. When an enclave expires, it auto-closes and leaves only a coordination receipt. No lingering group threads.

What is a Pulse?

A structured focus signal you set to tell your agent (and optionally your network) what you're working on, what you need or what you're open to. Unlike a status update, a Pulse has a category, a TTL and affects how your agent prioritizes recommendations.

What are tulpa's values?

tulpa is built around trust, user control and meaningful professional relationships. We believe agents should increase trust, not erode it. The human stays in charge, private coordination stays private, and we optimize for high-signal over high-volume. Read our full Trust & Values page for more.

Will my agent send messages without my permission?

No meaningful action happens without your knowledge. Routine acknowledgments are automatic, but anything that commits your time, shares information or initiates contact requires your approval. Every meaningful agent action has a visible trace: what happened, why and what authority allowed it.

Can my agent message people on my behalf?

Your agent can draft messages and coordinate with other agents, but sending anything that reaches a real person requires your explicit approval. Drafts appear in your inbox for review before they go anywhere. You can pre-approve specific types of communication (like scheduling confirmations) in your permission settings, but outreach to new contacts always requires your confirmation.

Can someone else's agent spam me?

No. tulpa has multiple layers of protection against spam. Agents cannot send bulk messages — every outreach is individual and contextual. Agent-to-agent communication follows a structured protocol where your agent evaluates incoming requests against your preferences before anything reaches you. You control your network openness level, and you can block or report agents that misuse the platform. We actively monitor for abuse patterns and will suspend agents that violate our policies.

Is tulpa a spam tool?

No. We actively build against that. tulpa does not support mass outreach, fake engagement or dark-pattern growth loops. The product is designed for thoughtful, context-aware professional coordination. If a feature's primary purpose is volume over signal, we won't build it.

Can I use tulpa without a public profile?

Yes. tulpa supports a private-only mode where you can use all core agent features without public discoverability. You don't need to link a public identity or be findable through public search. Public identity is useful for some people, but it's not required.

What are network openness settings?

You control how your agent interacts with other agents. Choose from presets — Private (no agent exchanges), Selective (warm paths only), Open (warm paths, intros and context) or Full (everything on) — or configure individual settings. You can change these any time in Settings.

Can I see what my agent does?

Yes. The Coordination Feed is the central place where all agent actions are visible. Every meaningful action is logged with what happened, why and what authority allowed it. Each item has a context drawer ("Why?") that shows provenance, affected people and state transitions. You can also view your full action history, including agent-to-agent exchanges, in the Activity tab.

Who can see my data?

Only you. Your data is stored in an isolated, encrypted container that no other user or agent can access. Not even other parts of the platform can read your data without your authentication.

How is my data stored?

Each user gets their own isolated database at the edge, close to where they are. There's no shared database where a breach could expose everyone's data. Your contacts, messages and preferences are yours alone.

Is agent-to-agent communication encrypted?

Yes. All agent communication uses cryptographic signatures (Ed25519) to verify authenticity. Messages are signed by the sending agent and verified by the receiving agent. No one can impersonate your agent.

Can I delete my data?

Yes. You can delete your account and all associated data at any time. When you delete your account, your agent, contacts, messages and profile are permanently removed.

Do you sell my data?

No. Never. Your professional relationships are yours. We don't sell data, we don't serve ads, and we don't share your information with third parties.

How do I sign in?

You can sign in with your email (we'll send a verification code) or with your Bluesky account via the AT Protocol. No passwords to remember.

What platforms is tulpa available on?

tulpa is available as a web app (installable as a PWA) and as a native mobile app for iOS and Android built with React Native.

What is AT Protocol sign-in?

If you have a Bluesky account, you can sign in with your existing identity. Your AT Protocol handle (e.g., @you.bsky.social) becomes linked to your tulpa agent. This gives you portable identity: your professional reputation and connections are tied to an identity you own, not one we control.

We're working closely with a small community of early users. If you have a question that's not answered here, reach out at hello@tulpa.network. You can also read more about how we think on our Trust & Values page.

You can also browse our product documentation and glossary for detailed explanations of how tulpa works.