> ## Documentation Index
> Fetch the complete documentation index at: https://docs.nodalmerge.com/llms.txt
> Use this file to discover all available pages before exploring further.

# NodalMerge Studio

> Git for agent reasoning and execution — not Git for files. An agent-native collaborative workspace built on NodalMerge.

# NodalMerge Studio

**Git for agent reasoning and execution — not Git for files.**

NodalMerge Studio is an agent-native collaborative workspace platform built on
[NodalMerge](/). It provides agent orchestration, task
management, projection-based context, MCP integration, human review workflows, and a
VS Code Control Tower — all backed by a persistent, branchable, replayable artifact
graph.

Every step an agent takes produces a durable artifact in a DAG. Every artifact can be
inspected, branched, replayed, merged, and audited. The durable graph is the product;
agents are features of it.

## The artifact pipeline

```
Goal
  ↓
Plan               ← inspectable, reviewable, branchable
  ↓
Work Units         ← inspectable, has lineage, has scope
  ↓
Branch Changes     ← auditable, attributable, comparable
  ↓
Merge Proposal     ← reviewable, approvable, supersedable, replayable
  ↓
Approved State     ← committed, traceable back to the goal
```

1. **You define a goal.** A human describes what needs to be accomplished.
2. **An orchestrator agent plans it.** The orchestrator breaks the goal into work units and tasks.
3. **Worker agents execute tasks.** Each worker operates on its own branch, producing changes.
4. **Merge proposals are generated.** Workers submit structured proposals with full lineage.
5. **A proposal is reviewed and approved.** By default a human approves it — a goal can opt into agent or hybrid review instead. See [Trust & autonomy](/studio/concepts/trust-and-autonomy).
6. **State converges.** Approved changes land in the authoritative branch.

## Why teams use Studio

<CardGroup cols={2}>
  <Card title="Durable artifact graph" icon="diagram-project">
    Every plan, task, proposal, and decision is a persistent DAG node. Trace any
    approved change back through its entire lineage to the original goal.
  </Card>

  <Card title="Human-governed by default" icon="user-check">
    Agents propose; humans decide. The merge review workflow enforces human approval
    before speculative work lands in the authoritative branch — unless a goal
    explicitly opts out.
  </Card>

  <Card title="Branch and replay" icon="code-branch">
    Branch from any artifact checkpoint to spawn alternate executions. Replay a
    proposal under a different model to compare outcomes. Roll back to any Known
    Good State.
  </Card>

  <Card title="MCP-native" icon="plug">
    A frozen MCP v1 tool surface with 66 tools across 19 namespaces. Integrate any
    MCP-compatible client with the Studio host.
  </Card>
</CardGroup>

## Get started

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/studio/quickstart">
    Install the Control Tower VS Code extension and run your first goal.
  </Card>

  <Card title="How Studio works" icon="brain" href="/studio/concepts/overview">
    The three-layer architecture and how Studio relates to the NodalMerge core engine.
  </Card>
</CardGroup>

## Scope today

Studio is v1. Deliberately out of scope for now: autonomous self-directed agents
(agents cannot set their own goals), long-term memory databases or vector stores,
cross-workspace reasoning, enterprise RBAC, and multi-tenant SaaS architecture. Agent-
controlled merges are possible but only when a goal explicitly opts into agent or
hybrid review — human approval remains the default.