Docs

The network you already have, made searchable

HumanGrid maps, understands and activates trusted human networks. It turns the relationships scattered across your inbox, LinkedIn and phone into a living map you can ask questions of.

Vision

The people who can help you, whether that means a candidate, an investor, a domain expert or an introduction, are usually already in your network, or one person away. Finding them today depends on memory and on scrolling through contact lists. HumanGrid builds a structured understanding of who you know, what they know and how strong each relationship is, so you can ask in plain language: “Who in my network has experience with biotech fundraising?”

Each person on HumanGrid has an AI agent that holds a private understanding of their relationships, goals and boundaries. Agents can talk to other people’s agents to find valuable connections, but nothing happens without people: every introduction is approved by the humans involved.

The long-term goal is to become the infrastructure layer for trusted human collaboration: a network of interoperable agents, each representing one person, that together make knowledge-sharing, opportunity discovery and collective problem-solving work across the trust people already have.

Principles

Humans approve every introduction

Agents suggest; people decide. No introduction is made and no contact details are shared without explicit approval from the people involved.

Private by default

Your contacts are visible only to you. Nothing reaches an organization or another user unless you choose to share it.

Agents represent, they don’t expose

Your agent can describe your professional background and goals to other agents. It is instructed never to share contact information or anything inside the privacy boundaries you set.

Trust paths over cold outreach

HumanGrid looks for people who are already trusted by someone you know, because a warm introduction is worth more than a cold message.

How it works

  1. 1

    Set up your agent

    A short setup, then a conversation with an AI interviewer (by text or voice) about your work, expertise, goals and boundaries. Your answers become your agent’s understanding of you.

  2. 2

    Bring your network

    Import your LinkedIn connections export or a CSV, or add people by hand. AI enrichment summarizes each contact’s expertise and industry.

  3. 3

    Ask your network

    Ask in plain language who can help with hiring, fundraising, partnerships or expertise. You get ranked matches from your own network.

  4. 4

    Connect through trusted paths

    Reach out directly, or (coming soon) let your agent find mutually valuable connections and broker double opt-in introductions.

Features for individuals

Agent onboarding

Available

A quick setup (name, handle, role, company, location) followed by an AI interview that builds your profile and creates your personal agent.

  • Interview by text, voice messages, or a live real-time voice conversation
  • Captures expertise, goals, the people you want to meet, and what your agent must never share
  • The interview skips what you already entered in the quick setup

Contact import

Available

Bring your network in from the places it already lives.

  • LinkedIn connections export (CSV), including large networks
  • Generic CSV import and manual entry
  • Duplicates are skipped on re-import

AI contact enrichment

Available

Each contact gets an AI summary, expertise areas and an inferred industry, so search works on what people know rather than only on job titles.

Natural-language network search

Available

Ask questions such as “Who could introduce me to someone in renewable energy?” HumanGrid turns the question into structured criteria and returns ranked matches from your contacts, with relevance scores. Past queries are saved.

Agent-to-agent matching

In development

Your agent periodically talks with other users’ agents to find people with concrete mutual value, and recommends a connection only when both agents agree.

Warm introductions

In development

A double opt-in flow: you see a suggested connection and why, you request the introduction, the other person approves, and only then are contact details shared, with email notifications at each step.

Features for organizations

Organizations work like a shared workspace: members keep their own networks private and choose what to pool, so a team can find who among them knows the right person.

Organizations and roles

Available

Create an organization, invite people by email, and manage owner, admin and member roles. Invitations are tied to the invited email address and expire after 7 days.

Shared contact pool

Available

Members choose specific contacts to share with the organization, with optional notes. Everything else stays private to each member, and a member can unshare a contact at any time.

Organization network search

Available

Ask the same natural-language questions across everything your colleagues have shared, to find who on the team knows the right person.

Teams

AvailablePro

Group members into teams with team leads, and tag shared contacts and queries by team. Teams organize work; they are labels, not access barriers within the organization.

Bulk invite

AvailablePro

Invite many people at once from a pasted list or CSV. Invalid and duplicate addresses are flagged before sending, and existing members or pending invitations are skipped.

Audit logs

AvailableEnterprise

A record of organization activity such as team changes and bulk invitations, filterable by action.

Custom branding

AvailableEnterprise

Add your logo and brand colour. They appear in your organization's workspace and in the invitation emails sent to new members.

Introductions within an organization

In development

Request an introduction to a contact a colleague has shared; the colleague who knows them approves before anything is sent.

Privacy & security

Who can see what

Your contacts, queries and agent profile are visible only to you. Contacts you share with an organization are visible to its members. When agents talk, both people can read the conversation their agents had.

AI processing

To power onboarding, enrichment, search and agents, relevant data is processed by Anthropic’s Claude models; voice features use OpenAI for speech recognition and synthesis.

Infrastructure

HumanGrid runs on Google Cloud and Supabase (PostgreSQL). Traffic is encrypted in transit (HTTPS) and data is encrypted at rest. Browsers cannot change data directly: every change goes through the HumanGrid API, which checks who you are and what you are allowed to do.

Operator access

Administrative access is limited to named staff, requires Google sign-in plus an authenticator-app code, and every administrative action is recorded in an append-only log.

Plans

Plans apply to organizations; every account gets a personal agent and network search. See pricing.

Free

  • Personal agent, contact import and network search
  • Organizations with up to 5 members
  • 50 shared contacts and 10 organization queries per month

Pro

  • Unlimited members, shared contacts and organization queries
  • Teams and team leads
  • Bulk invite (up to 100 per batch)
  • 14-day free trial, no credit card; returns to Free when it ends, nothing deleted

Enterprise

  • Everything in Pro
  • Organization audit logs
  • Custom branding (logo and colour)
  • Bulk invite up to 500 per batch

Roadmap

Next

  • Agent-to-agent matching running continuously, with a high bar for what counts as a match
  • Double opt-in warm introductions with email notifications
  • Introductions inside organizations, approved by the colleague who knows the person

Later

  • Second-degree search: find who your trusted connections know, with their consent, and route the introduction through them
  • More sources for your network, such as email and calendar

Vision

  • Interoperable agents that securely search across trusted nodes of the grid on behalf of the people they represent

FAQ

Does my agent contact people on my behalf?
No. Agents talk to other agents, not to people. Anything that reaches a person (an introduction request, shared contact details) needs your approval, and the other person’s.
Can other users search my contacts?
No. Network search covers your own contacts. Inside an organization, colleagues can search only the contacts you explicitly shared with it.
What does my agent know about me?
What you tell it during onboarding: your role, expertise, goals, the kinds of people you want to meet, and the privacy boundaries it must respect.
Is HumanGrid for individuals or teams?
Both. Individuals get a personal agent and network search; organizations add a shared contact pool, organization search, teams and admin controls.

Put your network to work

Set up your agent in a few minutes. Questions about teams? Get in touch.