Create an intent
People search for plans, peers, collaborators, activities or help using plain language.
B2B intent-based networking
LMGram helps expo, summit and conference attendees say what they need and get matched with people looking for the same thing, or the complementary side. Less cold networking, more qualified conversations.
Expo floor
Buyers and vendors matched around a real expo need.
Sarah We supply compostable bottles for small runs.
Michael I am comparing MOQs and lead times.
Priya We can discuss food-safe labeling options.
What LMGram does
Users express what they want in natural language. LMGram understands semantic overlap, applies location when useful, creates match requests and opens the right conversation surface: one-to-one chat or an intent room for a group.
People search for plans, peers, collaborators, activities or help using plain language.
The engine compares sessions by semantic similarity, shared terms, complementarity and distance.
Matches can become direct conversations with text, photos, video, audio and push notifications.
Developers and AI agents can create intents and inspect matches with lmgram-cli.
Clusters of compatible matches can become a group room so participants coordinate together.
How it works
New: lmgram-cli
AI agents can write code, analyze files, run tests and produce plans. But when the task needs a real person - an expert, beta tester, supplier, buyer, reviewer or founder - LMGram gives the agent a controlled path to create an intent and find matching humans.
$ lmgram login
Approve this agent in LMGram mobile
$ lmgram intent create "Need a founder who sold B2B SaaS into US SMBs"
Intent created: intent_8x41
Expires: 72h
$ lmgram matches list --intent intent_8x41
3 qualified matches
$ lmgram match accept match_24b
Chat opened
The CLI uses explicit LMGram authorization. Actions are tied to a real account and scoped credentials, not anonymous automation.
Agents can publish a concise human need exactly when they detect a blocker, then check matches without leaving the workflow.
Matches can become direct chats or intent rooms, so the user moves from agent workflow to real conversation with context preserved.
Intent rooms
Many real-world intentions are group-shaped: buyer/vendor discovery, founder roundtables, investor conversations, workshops, side events, dinners or teams looking for the same discussion. Intent rooms turn match clusters into a shared place to coordinate.
LMGram avoids duplicate group chats by recognizing when matched intentions belong to the same open cluster.
New people can join if they match any member's intent in an open room, even after the room was created.
Members can close a room to new matches when the table, team, group or activity is full.
Best-fit markets
Help attendees find peers, investors, customers, founders and side-event groups before the opportunity is gone.
Activate guests around dinners, sports, tours, business networking and local plans in the same building.
Connect buyers, suppliers, sponsors and exhibitors by explicit commercial needs instead of static directories.
Surface founders, builders, operators and freelancers who are looking for feedback, talent or collaboration.
Let Codex, Claude Code, Cursor, Windsurf and local agents route technical or commercial blockers to real people.
Connect builders looking for reviewers, beta testers, implementation advice, vendors or collaborators.
Trust and control
LMGram now includes safety and conversation controls across individual and group chats. These are foundational for hotels, campuses, tourism and any venue-based deployment.
Mobile, ChatGPT, CLI and API