Skip to main content

Open ACE - Self-Hosted Control Plane for AI Coding Agents

ACE = AI Computing Explorer

This guide explains what Open ACE is, which teams it is built for, and which parts of the product matter most when you evaluate it.

Project Positioning​

Open ACE is an open-source, self-hosted control plane for AI coding agents.

It brings three layers together:

  1. A browser workspace where developers run AI coding sessions
  2. A remote execution layer for running agents on internal machines
  3. A governance plane for keys, quotas, audit, compliance, and cost visibility

The project is aimed at teams that want more than a single-chat assistant. It is for organizations moving AI coding into real engineering workflows, where issues, repositories, credentials, remote machines, and reviewability all matter.

What Problems It Solves​

Team problemWhat Open ACE provides
Multiple AI coding tools are used in parallelOne workspace and control plane for Claude Code, Qwen Code, Codex, OpenClaw, and ZCode
Agents need to run on internal, staging, or GPU machinesRemote Agent executes CLI tools on controlled Linux, macOS, and Windows machines
API keys should not be copied across laptops and remote boxesAPI Key Proxy keeps real keys on the server and issues short-lived, revocable proxy tokens
AI work needs to be inspectable after the factAutonomous workflow timelines, milestone summaries, final code changes, and run provenance keep execution visible
Platform and security teams need governanceQuotas, anomalies, audit logs, compliance checks, multi-tenant controls, and SSO are built into the same system

Core Product Layers​

1. Autonomous Development Workflows​

Open ACE can turn GitHub issues into structured AI development runs.

Core workflow capabilities include:

  • Issue-driven runs with preparation, planning, development, review, reporting, and merge phases
  • Batch issue handling for operating on more than one GitHub issue from a single workflow
  • Pause, resume, retry, cancel, and auto-merge controls
  • Fork-from-here behavior at the milestone level
  • Final code-change summaries and milestone-level diff inspection
  • Timeline views that expose outputs, status changes, review results, and token usage

This makes Open ACE more than a browser wrapper around an AI CLI. It becomes an operational surface for autonomous or semi-autonomous repository work.

2. Workspace and Remote Execution​

Open ACE provides a browser-based workspace for both local and remote AI coding sessions.

Work Mode​

Work Mode is the day-to-day surface for developers:

  • Browser workspace for AI coding sessions
  • Session history and recovery
  • Prompt library and reuse
  • Multiple workspace tabs
  • Conversation and execution visibility

Remote Agent​

Remote Agent lets teams run AI coding tools on controlled machines instead of pushing all work onto a local laptop.

Key capabilities include:

  • Token-based machine registration and remote identity management
  • CLI adapters with session resume and permission-mode handling
  • Browser terminal access
  • Remote file and project workflows
  • code-server / VSCode proxy access for in-browser development
  • Run Timeline APIs for persisted remote execution provenance

3. Governance and Administration​

Open ACE also acts as a management plane for AI engineering usage.

Manage Mode​

Manage Mode is for platform, IT, and governance teams:

  • Usage dashboards and trend analysis
  • Quota management and quota alerts
  • Anomaly detection
  • Audit Center and Security Center
  • Compliance and retention workflows
  • Tenant, user, and permission management
  • ROI views with visible planning assumptions

Security and Access​

  • API Key Proxy with encrypted server-side key storage
  • Short-lived, revocable proxy tokens for local and remote sessions
  • OIDC / OAuth2 / SAML 2.0 SSO support
  • Multi-tenant organization model
  • Content filtering and audit logging
  • Feishu and DingTalk integration paths for org sync and alerts

Supported AI Coding Tools​

ToolSupport in Open ACENotes
Claude CodeYesWorkspace sessions, remote execution, permission modes, session recovery
Qwen CodeYesWorkspace sessions, remote execution, session recovery, usage integration
CodexYesWorkspace sessions and remote execution through CLI adapters
OpenClawYesWorkspace support plus session and message synchronization
ZCodeYesRemote Agent support, session sync, and persistent app-server execution mode

Why Teams Evaluate It​

Open ACE is especially relevant when a team needs some combination of the following:

  • Self-hosted deployment inside a private network
  • Shared access to multiple AI coding agents
  • Remote execution on development, staging, or GPU machines
  • Centralized LLM credential handling
  • Explainable autonomous runs instead of opaque background automation
  • Governance visibility across token usage, cost, quotas, anomalies, and audits

Deployment and Integration Story​

Open ACE supports several deployment paths:

  • Docker Compose quick start
  • Source installation for development
  • Package-based install workflows
  • Kubernetes reference deployment
  • Reverse proxy deployment with Nginx

The stack uses:

  • Backend: Python 3.10+, Flask, SQLAlchemy, Alembic
  • Frontend: React 18, TypeScript, Vite, Bootstrap 5
  • Database: SQLite and PostgreSQL
  • Remote execution: Remote Agent with CLI adapters, terminal relay, and session sync

Who Should Read What Next​

Start with the guide that matches your evaluation path:

Bottom Line​

Open ACE is not just a usage dashboard and not just a browser shell around an AI CLI.

It is a self-hosted control plane for teams that want to:

  • run multiple AI coding agents in one place,
  • execute them on controlled machines,
  • automate issue-driven engineering work,
  • and keep the whole process observable, governable, and reviewable.