AI-assisted personal network management

Your network.
Active. Yours. Alive.

A private workspace for the people you already know. Bring your contacts together, find the right person and make room for a meaningful follow-up. Optional local AI helps you see the connections.

See your network differently → See how it works

For You

Import your contacts.
Own your network.

Start with your existing relationships. Import a CSV, add context and explore your network through search, a spatial mesh and expertise clusters.

Import from

LinkedIn Google Spreadsheet
Illustrative contacts Fictional example
SA
Alex R.
Product designer · climate tools
core contact
JK
Jordan K.
Engineer · creative coding
imported
MP
Sam P.
Researcher · learning
core contact
LM
Riley M.
Founder · community spaces
imported
01 · Import
Bring your contacts in
Upload a CSV, connect LinkedIn, or sync Google Contacts. Your data stays on your device — never on our servers.
02 · Own
Your mesh, your device
An encrypted local vault holds your contact data. Add notes and mark your core relationships without uploading an address book to a hosted service.
03 · Explore
Find the connection
Search by name, role, company, location, tags or notes. Move between Mesh and Cluster views to understand the people around a topic.
04 · Reconnect
A thoughtful next step
Explore a small group around an idea and prepare an invitation. Review the draft, then send it yourself through the tools you already use.
See the current prototype →

An assistant, at your invitation

See the context.
Choose the next move.

AI can help you explore your network without taking over your relationships. Search, tags and the mesh work on their own. Connect a local model when you want another perspective.

A network around an idea Illustrative square nodes connect a central idea to design, research and community contacts. This is a visual explanation, not a screenshot or real contact data. YOUR IDEADESIGNRESEARCHCOMMUNITYENGINEERING
Illustrative network · fictional relationships
Local analysisAn optional Ollama model helps explore selected context from your own contacts.
Useful contextRoles, interests and your notes make a connection easier to understand.
Draft, then decideCluster suggestions and invitation drafts give you a starting point. You choose what to share.
Agent access is opt-inThe desktop prototype includes agent settings and local skill definitions. A working MCP connection is planned; it is not available in this build.

For your network

From a name you saved
to a conversation that matters.

A contact list remembers a name. InterMesh helps you remember the context and decide where a conversation could go next.

01
Keep the context
Collect contacts, add notes and mark the people at the core of your network.
02
Explore an idea
Look across expertise and interests. Suggest a small group from the contacts you already know.
03
Make it personal
Prepare an invitation, adjust it in your own voice and send it yourself. Nothing is sent automatically.

The current prototype

Your network.
On your terms.

InterMesh is a working desktop prototype in active development. This website introduces the product; it does not ask you to upload contacts.

Built today

CSV import, local search, contact notes, a core network, Mesh and Cluster views, cluster suggestions and invitation drafts.

Local vault

The desktop build stores sensitive fields in an encrypted SQLite vault. A browser preview uses a separate encrypted local vault. These are prototype implementations, not a claim of an independent security audit.

Optional AI

Local analysis uses Ollama when configured. Import, search, graph layout and cluster ranking work without an AI model. The agent connection remains a settings scaffold in this build.

You send

Invitations remain drafts. You review them and decide whether, where and when to send them.

Availability

A public desktop download is not available yet. Ask for a walkthrough of the current prototype. The earlier hosted experiment remains separate from this desktop direction.

Ask for a walkthrough →Earlier web experiment →