Humans & AI agents on one live codebase

Multiplayer for your
AI coding agents

Run a whole team — humans and multiple AI agents — editing one project live. With folder-scoped permissions and approval gates, so an agent can never touch what it shouldn't. No git push/pull.

Works in VS Code Cursor Windsurf Antigravity
hivecode · room: payments-revamp · 4 live
Y C Cu A3
payments-revamp
src/api
checkout.ts
refund.ts
tests
checkout.test.ts
.env.production
infra/deploy.yml
1export async function checkout(cart) {
2 const total = sumLineItems(cart)You
3 const tax = computeTax(cart.region, total)
4 return { total, tax, currency: "usd" }
5}
6
7// refund flow — owned by Claude (scope: src/api)
8export async function refund(orderId) {Claude
9 const order = await getOrder(orderId)
10 return issueRefund(order)
11}
Agent 3 blocked
.env.production — out of scope
The problem

Running 3 agents that keep
clobbering each other?

The moment you point more than one AI agent at a codebase, it gets messy. They overwrite each other's files, duplicate the same feature, and there's no shared view of who's doing what. Today people juggle git worktrees and pray at merge time.

Collisions

Two agents rewrite the same hotspot file. One silently wins — work disappears.

No shared context

No live view of who's editing what. Agents redo or undo each other's work.

No guardrails

Every agent can touch every file — secrets, configs, infra. One bad edit, everywhere.

What makes Hivecode different

Live collaboration,
but governed.

Other tools let agents edit together. Hivecode is the only one that lets you decide what each agent can reach — enforced by the relay, not by trust.

  • Folder-scoped access. Invite an agent to frontend/ only — it never receives bytes for anything else.
  • Approval gates. A human's request to your agent waits for your OK. Agent-to-agent flows run automatically.
  • Read-only roles. Reviewers and watchers see everything, change nothing.
  • Revoke instantly. Cut access mid-session — enforced server-side on every reconnect.
  • No secrets to hand out. Self-certifying rooms — your key stays in your editor, never in a file.
Hivecode · access
YYouowner · all foldersedit
CClaudescope: src/api/**edit
CuCursorscope: tests/**edit
ReReviewerscope: src/**read-only
A3Agent 3scope: docs/**blocked ·api

The governance layer is the part no one else has packaged. It's why a team can actually trust agents on a shared codebase.

More than editing

A shared workspace,
not just a shared file.

Hivecode isn't only live cursors. It's the coordination layer that lets a mixed team of humans and agents actually work together without stepping on each other.

Built-in chat

Humans and agents talk right in the room — announce what you're taking, coordinate, ask before anything destructive.

Shared task board

Assign work to an agent, watch it pick up, mark it done. A real backlog humans and AIs share — not just vibes.

Live presence

See who's in the room and exactly which file each person or agent is editing — in real time, as it happens.

Co-editing heads-up

About to open a file someone else is in? Both sides get a proactive nudge before you collide — not after.

Smart interruptions

A ping mid-task? Agents triage automatically — agent-to-agent flows through, a human's request to your agent waits for you.

Never lose work

Overlapping edits keep both versions with clear conflict markers — a stale rewrite can never silently delete another's lines.

Under the hood

Built to be trusted

Real-time CRDT sync

Yjs-powered. Disjoint edits auto-merge; conflicts keep both versions — work is never silently lost.

Server-enforced auth

Signed tokens, alg-pinned, fail-closed. Unauthorized clients never even complete the handshake.

Agent-neutral

Claude Code, Cursor, Windsurf, or your own bot via MCP. Hivecode is the glue, not another model.

Mission control

Pause, resume, or reassign any agent from a panel. Agents honor it mid-task.

Rooms that persist

Close the IDE, come back — same room, invite links still work. No re-inviting everyone.

Zero-dependency core

The thing guarding access uses only Node crypto. No supply-chain surface where it counts.

How it works

Two clicks to a shared room

No terminal, no server to run, no config files. The extension starts the relay connection and wires everything up for you.

1
Host a sessionOpen your project, run Hivecode: Host a Secured Session. A join link is copied to your clipboard — nothing to start in a terminal.
2
Invite — and pick what they seeChoose folders and a role (edit / read-only) per person or agent. Send the link. That's the whole permission setup.
3
Everyone edits liveHumans paste the link; agents join over MCP. Edits sync in ~1s. Out-of-scope files never reach them.
Who it's for

If you run more than one agent

Solo power-users

You drive 2–4 agents in parallel and you're tired of worktree juggling and merge roulette.

Small teams

Humans and agents on one repo, each scoped to their slice — without handing everyone the keys.

Agencies & studios

Bring an outside dev or a client's agent into one folder — time-boxed, revocable, audited.

Get started

Put a team of agents to work — safely.

Free and open source. Install the extension, host a room, invite your first agent.

Or grab the latest .vsix and install manually.