Free and Open Source — Works with Claude Code and OpenAI Codex

AI Agents That
Manage AI Agents.

Crest adds a layer of AI on top of your CLI coding agents. In-app agents orchestrate Claude Code and Codex—decomposing tasks, managing context, and recovering from errors—so multiple terminals work in parallel without you babysitting any of them.

Download for macOS Linux / Chromebook

Chromebook users: Enable Linux (Settings → Developers → Linux) first

View on GitHub
2 CLI Providers
25 Skills
5 Agents
14 Themes
Crest
Claude Code and OpenAI Codex running side by side in Crest

Three Layers of AI. One App.

CLI agents are powerful but need babysitting. Crest stacks AI on top of AI so you don't have to.

1

CLI Agents

Claude Code and OpenAI Codex run in real terminals inside Crest. They read your code, write files, run commands, and use tools. This is the layer that does the work.

2

In-App Agents

Crest's Super Agent watches each terminal, approves tool calls, answers prompts, manages context windows, and recovers from crashes. AI managing AI—so you don't sit and type "y" all day.

3

The Orchestrator

Give one big task. The Orchestrator breaks it into sub-tasks, spins up multiple terminals, assigns each a CLI agent with its own in-app agent, and coordinates them all. AI coordinating AI coordinating AI.

Built for Builders

Claude Code and Codex are incredible—but the terminal is a wall for most people. Crest tears it down. Whether you write code daily or have never opened a terminal, this is your way in.

For Developers

You already live in the terminal. Crest gives you the orchestration layer your CLI agents are missing.

  • Switch between Claude and Codex per tab—use the right model for every task
  • Dispatch the Orchestrator to break complex work into parallel agent streams
  • Monitor cost, token usage, and session history from a single analytics view
  • Built-in git workflow, file explorer, browser preview, and 25+ extensible skills
See developer use cases

For Everyone

No command line experience needed. Describe what you want to build and let Crest's agents handle the rest.

  • Super Agent runs autonomously with an outer LLM monitoring quality and progress
  • Zero config setup—open the app and start building immediately
  • See live previews of your project without touching a terminal
  • Teach the AI how you like things done and it remembers for next time
See non-developer use cases

Built for the Way Agents Actually Work.

Crest is not another wrapper. It is the orchestration layer, analytics dashboard, and coordination hub your AI workflow is missing.

Super Agent

Layer 2 AI: an in-app agent watches your CLI agent's terminal output, auto-approves tool calls, answers prompts, manages context, and recovers from errors. You set a task and walk away—the AI manages the AI.

Orchestrator

Layer 3 AI: give one complex task, the orchestrator decomposes it into sub-tasks and dispatches them across multiple terminals running separate CLI agents. Each terminal gets its own in-app agent managing it. AI coordinating AI coordinating AI.

Agent Hive (Swarms)

Launch specialized agent swarms—security audit, refactoring, test generation—all running simultaneously across your codebase. Each swarm agent is a CLI agent managed by an in-app agent, all coordinated from one dashboard.

Project Memory

Natural language memory that writes CLAUDE.md files. Tell it your preferences, architecture decisions, and conventions—it remembers across every session.

Teams Monitor

Watch Claude's multi-agent teams in real time. See team members, task assignments, progress, and coordinate across parallel workstreams.

Multi-Provider Terminals

Run Claude Code and Codex side by side. Each tab independently selects its CLI provider. Split panels, 3x2 grid, drag-and-drop tabs between panels.

Skills, Agents & MCP

25 skills and 5 agents ship with the app. One-click MCP server installs. Create, edit, duplicate, and share your own skills and agents.

Analytics Dashboard

Track sessions, tokens, coding hours, and project activity. 7-day and 30-day views, per-project time tracking, and activity charts.

Conversation History

Every conversation saved, searchable, resumable. Pin important sessions. Pick up exactly where you left off, even weeks later.

See all 50+ features

See It In Action

Explore every corner of the Crest experience

Terminal
Terminal with Claude Code, file explorer, and status bar File Explorer Terminal Tabs Model Selector Session Stats

Full terminal with file explorer, model selector, plan/work toggle, and session tracking

Multi-Provider Terminals
Claude Code and Codex running side by side Claude Code OpenAI Codex Per-Tab Provider

Run Claude Code and OpenAI Codex side by side—each tab picks its own CLI provider

Orchestrator
Orchestrator with split and parallel modes Split / Parallel Modes Task Assignment Terminal Count

Split or parallel mode—assign tasks per terminal and launch coordinated multi-agent workflows

Claude's Memory
Natural language memory with project and global scopes Scope Selector Natural Language Input Stored Memories

Teach Claude your preferences in plain English—scoped per project, private, or global

Teams
Teams monitor with task progress Active Teams Team Members Task Progress

Monitor multi-agent teams with real-time task progress and coordination

Analytics Dashboard
Analytics Dashboard Usage Stats Time Period Activity Chart Top Projects

Track coding time, tokens, and project activity with real-time analytics

Agent Hive
Agent Hive - multi-agent swarm management Hive Presets Agent Cards Launch Controls

Manage specialized agent swarms for audit, action, design, and more

Super Agent
Super Agent - autonomous task execution Task Input Time Limit Safety Level

Launch autonomous tasks with time limits, safety levels, and provider choice

Skills & Agents
Skills and agents marketplace Category Filters Search Skill Cards One-Click Install

Browse 25 skills and 5 agents—search, filter, and deploy in one click

What You Can Build

If you can describe it, AI can build it. Here's what people are using Crest for.

Full-Stack Apps

Go from idea to deployed app. Crest handles the code, tests, git, and deployment—you just describe what you want.

Security Audits

Deploy agent swarms to scan your codebase. Find vulnerabilities, fix them, and verify the fixes—all in one session.

Automate Anything

Scrape websites, process spreadsheets, generate reports, build dashboards. If you can describe it, Claude can build it.

Powered By

Anthropic
OpenAI
React
TypeScript
Tailwind
Electron

Stop Prompting. Start Orchestrating.

Download Crest for free and put every AI coding agent you use under one roof. Open source. Zero config. Immediate results.

Download Free View on GitHub

Works on macOS, Linux, and Chromebook

Download Crest