GitHub: https://github.com/mertkayacs/reevesagents

ReevesAgents is a free and open source workspace manager for AI CLI agents. It helps you use several vendor CLIs at the same time, without turning your local machine into a heavy platform or locking the work into one model.

It is built for people who want the best parts of multiple providers while keeping the workflow economical. Put each CLI where it is strongest: ask DeepSeek to work on backend code, use Claude for product and website direction, use Codex for a design system or implementation pass, and keep Hermes agents on mail, web, search, or research tasks when those CLIs are configured for that work. ReevesAgents gives the work a shared local shape: runs, agents, tmux windows, history, a TUI, and a local Web UI.

The UI is available in 9 languages: English, German, French, Spanish, Portuguese, Italian, Turkish, Simplified Chinese, and Arabic.

Use more than one CLI without losing the thread. If you already jump between Claude, Codex, DeepSeek, Hermes, OpenCode, or other agent CLIs, ReevesAgents puts those sessions into one local workspace.

Stay vendor-flexible. Provider login still belongs to each provider CLI, but ReevesAgents does not become another place for credentials or model traffic. You can add, remove, or switch CLIs without rebuilding the way you manage runs.

Keep cost practical. Route work to the CLI that makes sense for the task instead of pushing everything through one expensive default.

See the work at a glance. The Web UI is especially useful when several agents are running: active runs, agents, models, permission modes, stop/delete actions, and history are visible from one browser view while tmux keeps the real CLIs alive.

This launch-week demo shows the core workflow: split one project across the CLIs that fit each job, keep every session local in tmux, and watch the whole run from the TUI or Web UI.

Recorded with VHS. Pick a language, jump from the ASCII-art welcome screen into active runs, then open the run to see agents, tasks, status, and history.

The same local run in the browser: DeepSeek drafts the backend, Claude Code works on product and web direction, Codex reviews the implementation path, and Hermes waits for explicit research approval.

npm install -g reevesagents@1.2.0
reevesagents doctor
reevesagents

One-shot:

npx reevesagents@1.2.0 doctor

pnpm add -g reevesagents@1.2.0
reevesagents doctor
reevesagents

One-shot:

pnpm dlx reevesagents@1.2.0 doctor

yarn dlx reevesagents@1.2.0 doctor

Yarn Classic global install:

yarn global add reevesagents@1.2.0
reevesagents doctor
reevesagents

bun add -g reevesagents@1.2.0
reevesagents doctor
reevesagents

One-shot:

bunx reevesagents@1.2.0 doctor

brew tap mertkayacs/reevesagents
brew install reevesagents
reevesagents doctor
reevesagents

Use source when you want to inspect the code, contribute, or test the pre-beta research package from the repository.

git clone https://github.com/mertkayacs/reevesagents.git
cd reevesagents
pnpm install
pnpm build
pnpm link --global
reevesagents doctor
reevesagents

ReevesAgents is local-first. It expects a normal developer machine with tmux and at least one provider CLI already installed.

• macOS, Linux, or WSL.
• Node.js 20.19+.
• tmux. Version 3.0+ is recommended.
• A normal interactive shell on PATH.
• At least one supported provider CLI on PATH.

Native Windows is not the target runtime. Use WSL for Windows machines.

ReevesAgents can launch these provider CLIs when they are installed and authenticated on your machine:

• Claude Code
• Codex CLI
• OpenCode
• Hermes
• Kimi
• DeepSeek
• Pi
• Qwen
• Aider

Provider login, models, tools, quotas, and permission prompts stay with each provider. ReevesAgents does not store provider API keys and does not proxy model traffic.

Start the TUI:

reevesagents

Start a named run from the CLI:

reevesagents spawn deepseek:backend claude-code:product codex:system hermes:research --name "launch week build" --prompt "Plan the backend, product surface, design system, and research notes."

Open the local Web UI:

reevesagents web

Check your machine:

reevesagents doctor

Instead of running one model and hoping it is the right one for every task, you can split work by strength and cost. One run can hold a backend agent, a design agent, a research agent, and a review agent. Each one stays in its own tmux window, while the TUI and Web UI give you a calmer way to see what is alive, what ended, and what needs attention.

This is not a cloud agent platform. It is a small local layer around real CLIs.

The Web UI is local and loopback-only.

reevesagents web

It lets you create runs, add agents, choose provider models and permission modes, stop agents, delete ended work, and inspect history from the browser while the real CLIs keep running in tmux.

The Web UI uses optional runtime packages that npm installs by default. If you install with optional dependencies omitted, the CLI and TUI still work and the Web command explains what is missing.

The stable reevesagents package is the CLI, TUI, and Web UI for direct agent runs. MCP-connected root/child orchestration is separate and pre-beta.

Use this mode only if you want to test root/worker coordination, MCP setup, and prompt injection that teaches agents how to use the ReevesAgents registry.

Research-only disclaimer: orchestrator mode is experimental. Do not use it to create unattended fully autonomous agent systems. Be careful when routing outputs between different provider CLIs, especially if one provider output is used to prompt or evaluate another provider. Provider terms, data policies, and automation rules differ, and some combinations may violate a provider's rules. Review the terms for the CLIs you use and keep a human in control.

npm install -g reevesagents@1.2.0 reevesagents-orchestrator@pre-beta-research
reevesagents-orchestrator setup
reevesagents web --prebeta-orchestrator

Equivalent global installs:

pnpm add -g reevesagents@1.2.0 reevesagents-orchestrator@pre-beta-research
yarn global add reevesagents@1.2.0 reevesagents-orchestrator@pre-beta-research
bun add -g reevesagents@1.2.0 reevesagents-orchestrator@pre-beta-research

From source:

git clone https://github.com/mertkayacs/reevesagents.git
cd reevesagents
pnpm install
pnpm build
pnpm --dir packages/orchestrator build
pnpm link --global
pnpm --dir packages/orchestrator link --global
reevesagents-orchestrator setup
reevesagents web --prebeta-orchestrator

reevesagents-orchestrator setup may write MCP entries into provider CLI configuration. Run reevesagents doctor and reevesagents-orchestrator setup --help first if you want to review the path before changing local provider config.

You do not need:

• ReevesAgents-stored API keys.
• A database.
• Docker.
• A background service or daemon.
• MCP setup for normal stable agent runs.
• Provider config writes during the stable reevesagents install.

Install is passive. The stable package has no postinstall script and does not rewrite provider configuration.

End users do not need the development toolchain. Contributors use pnpm, TypeScript, tsup, Vitest, ESLint, and Prettier from the repository.

The duck art is already committed. Re-rendering brand assets is a maintainer task, not a user prerequisite.

Apache-2.0