Keep your work close.
Local-agent discovery and conversation continuity are the foundation. Find the agents on your device and keep the thread of your work in a client built around you.
Explore the local-first architectureAgent collaboration. Your way.
An open-source, local-first Human-Agent Collaboration Client. Bring your ideas to local AI agents, keep the conversation going, and stay in control of what happens next.
Less starting over.
More moving forward.
An idea becomes a project. One question leads to another. Give the conversation, the context, and your decisions a place to stay connected.
An idea with too many loose ends?
Start with a shared brief. Keep the goal, the decisions, and the next step in view.
Explore everyday use casesDifferent agents. Different ways in?
Understand how coding-agent interfaces handle sessions, context, and approvals before choosing your workflow.
Compare AI coding-agent interfacesWhat does your agent get to see?
Make context sharing and local actions deliberate, with clear privacy and approval boundaries.
Understand privacy and controlBring something
you care about.
A project to make, a question to explore, a task to pick up again. See how a conversation can give your next step a little more shape.
A shared starting point.
LicoUpNext: choose the first task together.
Next: follow the question that matters to you.
Next: continue with the context in view.
Illustrative conversations. Check current agent availability
Agent collaboration,
with you at the center.
You bring the intent. Agents contribute their capabilities. LicoUp is built around the conversation, context, and decisions that connect the work.
The next move is yours.
Choose what to share and what to allow. LicoUp keeps approvals explicit and asks for device permissions when your action needs them.
Read about context, privacy, and approvalsA good question
deserves a real answer.
Go a little deeper. Clear explanations for the way you want to work with agents.
What is agent collaboration?
People and AI agents working toward a shared goal. Learn how Human-Agent Collaboration keeps you involved, and where Multi-Agent Collaboration fits.
Meet the Human-Agent Collaboration ClientThe coding-agent landscape
Claude Code, Codex, Cursor, GitHub Copilot, and OpenCode: five representative interfaces and what they mean for collaboration.
Interfaces & compatibilityOrchestration or collaboration?
One organizes how work runs. The other makes participants, context, and decisions understandable. See how the layers fit.
A practical comparisonWhat makes a client local-first?
Follow the conversation model, agent adapters, and endpoint boundaries that shape LicoUp.
Inside the architectureOpen source.
Open possibilities.
Follow the code, understand the decisions, and help shape the next chapter of human-agent collaboration. LicoUp grows in the open.
Explore the sourceOpen source · AGPL-3.0-or-laterCheck current availability
A clearer starting point.
Straight answers about the product, your agents, and your control.
What is LicoUp?
LicoUp is an open-source, local-first Human-Agent Collaboration Client. It brings a human-facing conversation layer to agent collaboration, with local-agent conversations as its current foundation.
Explore the productHow do human-agent and multi-agent collaboration fit together?
Agent Collaboration is the broad category. Human-Agent Collaboration keeps a person actively involved in goals, context, and approvals. Multi-Agent Collaboration describes work involving several agents. A workflow can involve both.
Understand the collaboration modelWhich AI agents does LicoUp work with?
LicoUp includes built-in conversation adapters for 13 agent targets, including Claude Code, Codex, Cursor, GitHub Copilot, and OpenCode. They connect to locally installed agents through native CLI or local protocol interfaces. Actual compatibility depends on the agent version, available interface, and authentication setup.
Check current compatibilityCan I download LicoUp?
LicoUp v0.1.2 is available for macOS on Apple silicon. The public release includes the DMG and verification files. The status page brings together the current release, implementation, and roadmap.
See the macOS releaseDoes local-first mean everything works offline?
Local-first describes where the client and its conversation authority live. An agent you choose may still use an online model provider, which receives the content sent to it. Choose agents and context with that boundary in mind.
Understand the privacy boundariesWhere is LicoUp heading next?
LicoUp is growing from local-agent conversations toward a shared experience across people, agents, and devices. Peer protection is a preview capability; full Lico Arc integration and federation belong to the roadmap.
Follow current progressYour next idea.
A little more possible.
Download for macOSv0.1.2 · For Macs with Apple silicon