Skip to main content

Welcome to 4RGE

4RGE pairs a powerful coding agent with the controls a team needs to run it safely: a shared model catalog, usage and budget governance, project tracking, and cloud agents that keep working when you close your laptop.

Whether you are shipping a feature on your laptop, delegating a refactor to a cloud agent, or managing model access for a hundred engineers, 4RGE gives you one platform with consistent policy, auditability, and tooling.

This documentation is searchable — press ⌘K (or Ctrl K) anywhere to jump to any page or topic.

Production documentation URL

These docs are published at https://4rged.ai/docs. The site is built from MDX in apps/docs/content with navigation driven by _meta.json; run npm run validate:content in apps/docs to check manifests and frontmatter before contributing. Related product surfaces include 4rged.ai/agents (cloud workspace), 4rged.ai/projects (PM), and registry.4rged.ai (plugin marketplace index).

Start here

Platform capabilities

4RGE is more than a chat window in an editor. The platform spans the surfaces where engineers work and the controls that let organizations run agents at scale.

CapabilityWhat it gives you
Desktop agentAn IDE-integrated agent that reads your repo, edits files, runs commands, and asks for approval before sensitive actions
Cloud agentsLong-running agents in isolated cloud environments — continue after you close your laptop
ProjectsScoped work areas with tasks, views, and member roles tied to repositories
Org dashboardMember management, model catalog, budget policies, usage reporting, and org settings
Tool brokerMediated access to repository, terminal, browser, MCP, and workspace tools within policy
Plugin marketplaceExtend agents with MCP servers and integrations for external services
CLITerminal-first agent runs, scripted workflows, and headless CI handoff
ConsoleEnterprise operator controls for contracts, capacity, and cross-org governance
IntegrationsGitHub, Slack, Linear, and more for repo access, notifications, and automations

The desktop agent and cloud agents share the same account boundary, model catalog, and tool governance. A task you start locally can hand off to the cloud; a cloud run's diff lands in the same review workflow as local work.

Pick your entry point

Individual contributors usually start with the Quickstart and Agent overview. Team leads and admins should read Core concepts and the Dashboard overview. Platform operators start with the Console overview.

Who 4RGE is for

4RGE is built for software teams that want AI-assisted development without giving up control, visibility, or security.

Software engineers use the desktop app or CLI to describe changes in plain language, review agent-generated diffs, and approve tool runs. The agent respects your working directory, .gitignore, and the approval gates you set.

Engineering leads and tech leads organize work into projects, track tasks and progress, and delegate large refactors or test-fix sweeps to cloud agents so the team can keep shipping locally in parallel.

Organization admins manage members, invitations, and roles from the dashboard. They configure which models are available, set budget policies, and curate plugins so agents only reach the services the org approves.

Platform and IT operators (enterprise deployments) use the Console to manage contracts, capacity, egress policy, and governance across organizations — separate from day-to-day org admin work.

DevOps and automation owners wire 4RGE into CI with the CLI, connect GitHub for pull-request workflows, and set up cloud agent automations triggered by repository events.

Security overview

Safety is layered: user approvals, scoped tool access, organization policy, and — for enterprise customers — operator-level governance in the Console.

Approvals and scoping. The agent asks before running terminal commands or editing files outside the safe set. It operates within the folder you opened and honors .gitignore. Changes arrive as reviewable diffs you accept or discard. See Agent security.

Tool broker. Rather than giving agents unrestricted machine access, the tool broker mediates repository, terminal, browser, MCP, and workspace tools. Cloud and desktop runs flow through the same mediated layer, which keeps tool use auditable.

Organization controls. Admins restrict model and provider access in Model control, cap spend with budget policies, and manage member sessions. Cloud agent egress and network isolation are configured in Cloud agents · Security & network.

Secrets and credentials. Do not paste secrets into prompts. Use environment variables or your secret manager, and keep sensitive files out of the agent's reach via .gitignore.

Enterprise governance. Platform operators set contracts, capacity ceilings, and cross-org policy in the Console. Org admins set house rules inside those bounds.

Corporate networks

If you are behind a proxy or strict firewall, confirm the domains 4RGE needs to reach before rollout. See Cloud agents · Security & network and Troubleshooting.

Explore by area

How to read these docs

Search first

Almost every question — "how do I change my model?", "how do I invite a teammate?", "how do I set a budget?" — has a page. Use search (⌘K) and type what you want to do.

Code blocks are copy-pasteable; commands assume a POSIX shell unless noted. Internal links use paths like /getting-started/quickstart — you can paste them into the search bar or navigate from the sidebar.

Tables in these docs are resizable: drag column borders if content is truncated on smaller screens.

Troubleshooting

Setup issues, port conflicts, sign-in problems, or integration errors — see Troubleshooting.

For in-app help, deeplinks, and API references, see the Reference section.