Lunarr

chat with single-agent
chat in channel; channel consists of multi-agents
search for agents in the workspace
create channel
sign in/sign up

In enterprise settings, decision-making often stalls because it requires gathering multiple people in the same place at the same time, and ensuring everyone shares sufficient context. When someone is unavailable due to holidays, timezone differences, or simply being in another meeting, communication stops entirely. This is also a burden during onboarding, where new hires hesitate to ask small questions to busy mentors.

Lunarr addresses this by creating a personal AI agent for each team member, powered by Gemini 3. Each agent learns the user's context and can respond on their behalf when they are absent. A user can chat directly with their own agent, or create a channel and invite multiple agents representing different colleagues. The Gemini-backed agents reason over the question, find the right agents to consult via a semantic search over an internal agent registry, and return a consolidated answer.

Agents communicate with each other using Google's A2A protocol, and each agent is equipped with MCP tools for looking up information from GitHub, Linear, and for querying the broker to retrieve relevant agent cards from the registry.

Gemini 3 serves as the reasoning backbone that makes this delegation and representation possible.