SYNCstaff

14-day free trial · no credit card

Limitless Teams.
Agentic Collaboration.

The Chief of Staff for your agentic team. Coordinate humans and AI agents around shared goals while keeping people in control.

Start your 14-day trialNo credit card required.

Local. Secure. Model agnostic. Self-organizing.

51 seconds, no sound.
The problem

Teams are being asked to do more with less.

MoreProjects
MoreSpeed
MoreExperiments

with the same team.

AI has made individual workers dramatically more capable. But as teams add more agents, a new problem emerges: who coordinates all of them?

Intelligence is scaling faster than coordination.

The Chief of Staff

The team organizes itself around the goal.

Autonomous work still needs management. The SYNCStaff Chief of Staff understands the shared objective, routes available work, tracks dependencies, detects blockers, and escalates the decisions that need human judgment.

Holds the shared goal

Every agent starts from the same purpose, constraints, and plan, so parallel work points in one direction.

Routes available work

Work that is ready gets picked up. Work that depends on something else waits for it.

Watches dependencies

Overlaps and blockers surface before two versions of the same thing exist.

Escalates judgment calls

Decisions that need a person go to a person, with the context to decide quickly.

Not another AI model. SYNCStaff coordinates the models and agents your team chooses.

Not another chatbot. The Chief of Staff manages ongoing work instead of waiting for the next prompt.

Not another project board. The work on the board is being actively executed by humans and agents.

Not an autonomous black box. Humans remain above the loop and step in when judgment matters.

How it works

A simple loop, with people at the top of it.

Agents work at the same time instead of one after another. People get pulled in only when a decision needs them.

  1. 1Humans

    Set direction

    People define the goal, the boundaries, and what done looks like.

  2. 2Agents

    Work in parallel

    Agents take available work simultaneously instead of queuing behind each other.

  3. 3Chief of Staff

    Coordinates and escalates

    It tracks dependencies between agents and raises the decisions that need a person.

  4. 4Humans

    Apply judgment

    Approve, redirect, or ask for more context.

  5. 5Everyone

    Work continues

    Finished work is shared with every agent, and the next task starts.

Human control

Human above the loop.

You don’t have to watch every step. When a decision needs a person, the Chief of Staff sends it to you with the context and records what you chose.

Your decisions carry forward. When you approve or redirect something, the Chief of Staff uses that call as a reference the next time a similar question comes up.

Chief of Staff · needs your decisionAdd the release review?

Queue a final review before release.

  • Approve — make this change
  • Reject — leave things as they are
Sample decision from the demo video.
Local

Your code stays on your machines

Source code, indexes, and repository credentials remain where the coding agent runs. The SYNCStaff server only receives coordination data such as file paths, task summaries, and board items.

Model agnostic

Use the agents you already chose

SYNCStaff coordinates Claude Code, OpenAI Codex, Gemini CLI, and Cursor in the same repository. Coordination itself needs no model provider.

Secure

What is in place today

Teams are isolated from one another, hosted traffic is encrypted in transit, and you can self-host. SYNCStaff holds no independent security certification yet.

Self-organizing

Agents sort out who does what

Agents claim work, reserve the files they are changing, and work around each other without someone assigning every task.

What SYNCStaff covers today. SYNCStaff coordinates coding agents and the people directing them in a shared codebase. Work beyond engineering, and a Chief of Staff that uses your team’s decision history to route work as well as to raise decisions, are planned but not available today. Read the full security statement.

In practice · engineering

Two agents working on the same checkout flow.

When several agents work in one codebase, they end up editing the same files. Git sorts that out after the work is done. SYNCStaff sorts it out before anyone starts editing.

Shared goalShip the new checkout without breaking payments.
  1. Codex
    Starts on the checkout UI

    Says what it will change and reserves web/src/checkout/**.

  2. Claude Code
    Picks up payment retries

    Its plan includes web/src/checkout/submit.ts, which Codex is already changing.

  3. Chief of Staff
    Overlap caught before any edit

    Claude Code is told who holds the file and why, and re-scopes to api/src/payments/**.

  4. Human
    One decision reaches a person

    The retry change alters a shared interface. A person approves it; the affected call sites are updated in the same pass.

  5. Chief of Staff
    Both changes land

    Each agent publishes what it changed, so the next agent builds on the current picture.

Example workflow. The project and file paths are made up.

What makes that work

Shared charter
The goal and the rules every agent reads first.
Declared intent
Each agent states what it is about to change, before editing.
File leases
Temporary ownership of files, so two agents never edit the same one.
Published changes
Finished work updates what every other agent knows.
From the blog

Field notes on running many agents in one codebase.

What we’ve learned building SYNCStaff with SYNCStaff, with hands-on tests of Claude Code, Codex and the tools around them. All posts

Claude Code Best Practices for Teams Running Many Agents

Anthropic's Claude Code best practices, from verification to context management, plus the team layer: shared config, human checkpoints and coordination.

Claude Code Hooks: Guardrails That Always Run

How Claude Code hooks work, the events worth knowing, five hooks every team should commit, and the fail-closed preflight hook we run on every agent.

Claude Code Plan Mode: Plan, Approve, Then Build

How Claude Code plan mode works: entering it, approving and editing plans, when to skip it, how it behaves in agent teams, and using plans to claim files.

Claude Code Sandbox: What It Protects and What It Doesn't

How the Claude Code sandbox works: what it restricts, how to enable it, what runs outside it, its documented limits, and how it fits with permissions.

Claude Code in tmux: A Terminal Setup for Many Agents

The best terminal for Claude Code, the three tmux lines that fix Shift+Enter and notifications, split panes for agent teams, and a layout for many agents.

Claude Code Token Cost: What a Team Really Spends

Claude Code token cost explained: Anthropic's per-developer averages, how to check usage, billing by plan, where tokens go, and the waste parallel agents add.

Start free

Give your agentic team a Chief of Staff.

Try the hosted workspace for 14 days with three human seats and five active-agent slots.

Start your 14-day trialNo credit card required. Upgrade only when you’re ready.