Nullspace gives agents isolated compute, files, networking, terminals, and desktop environments. There are four integration paths — the right one depends on where the reasoning loop runs and whether Nullspace should operate a named project for you.

Local MCP

Your coding agent stays on your machine (Codex, Claude Code, Cursor, VS Code) and drives Nullspace through MCP tools.

Agent workspaces

The coding agent itself runs inside a machine — isolated filesystem, package installs, terminal, git, optional desktop.

Functions

Your agent project becomes a named, versioned job or service that your backend can call repeatedly.

Framework integrations

Nullspace as a tool inside your app (OpenAI Agents SDK, LangChain) — or your framework app deployed onto Nullspace (Claude Agent SDK, LangGraph, CrewAI).

Which path fits?

You want to…UseTradeoff to know
Let a local coding agent create machines, run commands, and fetch filesLocal MCPThe agent stays on your machine; only the work it launches runs isolated.
Put the coding agent itself in an isolated Linux workspaceCodex, Claude Code, Amp, or OpenCodeProvider credentials enter the machine environment.
Operate a repeatable agent app with runs, logs, outputs, and cleanupFunctionsYour backend owns product state and calls a named job or service.
Expose machine operations as one tool in a larger agent appOpenAI Agents SDK, LangChain, or bring your ownYour app owns retries, tool schemas, and framework state.
Run a framework app as the managed process itselfClaude Agent SDK, LangGraph, or CrewAINullspace owns runs and machines; the framework owns sessions and checkpoints.
Give an agent a browser or GUIDesktop environments
Let a human inspect the same machinePTY, SSH, or retained machines
Two notes that save confusion later:
  • Functions are jobs or services. A job is finite work with outputs. A service is a long-running HTTP process with URL, logs, and backing machine access.
  • Nullspace run IDs are not conversation sessions. Framework session, thread, checkpoint, and flow state stays app-owned. Use volumes, retained machines, hibernate/auto-resume, or an external database when state must outlive one machine process.

The object model

Whichever path you pick, the same technical components apply:
PrimitiveWhat it means
MachineThe isolated VM: commands, files, networking, PTY, SSH, desktop, volumes.
Agent projectA local directory with your app, a function.py definition, and an entrypoint.
FunctionA named, uploaded project runnable as a job or service.
Run / service instanceOne concrete execution, with logs, status, outputs, and a backing machine.
Framework stateSessions, threads, checkpoints, memory — owned by your app, not Nullspace.

Start with an example

Minimal job

Deploy and run a small project end to end.

HTTP service

A long-running Function service with a public URL.

Workspace checks

Clone a repo in a machine and run checks like an agent.