Paseo: Open-Source Interface for Claude Code, Codex, Copilot, OpenCode, and Pi Agents

✍️ OpenClawRadar📅 Published: June 4, 2026🔗 Source
Paseo: Open-Source Interface for Claude Code, Codex, Copilot, OpenCode, and Pi Agents
Ad

Paseo is an open-source coding agent interface that lets you run Claude Code, Codex, GitHub Copilot, OpenCode, and Pi agents from a single dashboard — on your phone, desktop, or CLI. It's self-hosted, privacy-first, and supports voice control.

How It Works

Paseo runs a local daemon that manages agent processes. Clients (desktop app, mobile app, web app, CLI) connect to it via WebSocket. Agents execute on your machine with your full dev environment, tools, configs, and skills.

Key Features

  • Multi-provider: Claude Code, Codex, Copilot, OpenCode, and Pi through the same interface. Pick the right model for each job.
  • Voice control: Dictate tasks or talk through problems hands-free.
  • Cross-device: iOS, Android, desktop, web, and CLI. Start at your desk, check in from your phone, script from the terminal.
  • Privacy-first: No telemetry, tracking, or forced log-ins.
  • Parallel agents: Run multiple agents concurrently on your own machines.

Getting Started

Prerequisites: at least one agent CLI installed and configured. The desktop app is recommended — download from paseo.sh/download or the GitHub releases page. Open the app and the daemon starts automatically. To connect from your phone, scan the QR code in Settings.

For CLI / headless setups:

npm install -g @getpaseo/cli
paseo

This shows a QR code in the terminal. Connect from any client.

CLI Commands (Reproduced from Source)

Run agents, list them, attach to live output, and send follow-ups:

paseo run --provider claude/opus-4.6 "implement user authentication"
paseo run --provider codex/gpt-5.4 --worktree feature-x "implement feature X"
paseo ls
paseo attach abc123
paseo send abc123 "also add tests"

Connect to a remote daemon with --host:

paseo --host workstation.local:6767 run "run the full test suite"
Ad

Skills (Agent Orchestration)

Install skills to teach agents to orchestrate other agents:

npx skills add getpaseo/paseo

Then use in any agent conversation:

  • /paseo-handoff — hand off work between agents (e.g., plan with Claude, handoff to Codex to implement).
  • /paseo-loop — loop an agent against acceptance criteria (Ralph loops), optionally with a verifier.
  • /paseo-advisor — spin up a single agent as an advisor for a second opinion.
  • /paseo-committee — form a committee of two contrasting agents for root cause analysis and planning.

Development

The monorepo includes packages:

  • packages/server — daemon (agent orchestration, WebSocket API, MCP server)
  • packages/app — Expo client (iOS, Android, web)
  • packages/cli — CLI for daemon and agent workflows
  • packages/desktop — Electron desktop app
  • packages/relay — remote connectivity
  • packages/website — marketing site and docs

Common commands:

npm run dev
npm run dev:server
npm run dev:app
npm run dev:desktop
npm run dev:website
npm run build:server
npm run typecheck

Self-Hosted Relay

For remote daemon access, use the self-hosted Go relay. Note: self-hosted relays use ws:// unless TLS is opted in. Configure with environment variables:

PASEO_RELAY_ENDPOINT=127.0.0.1:8080
PASEO_RELAY_PUBLIC_ENDPOINT=relay.example.com

📖 Read the full source: HN AI Agents

Ad

👀 See Also

Terminal-Based 3D Renderer Built with Multi-Agent Claude Code System
Tools

Terminal-Based 3D Renderer Built with Multi-Agent Claude Code System

A developer created tortuise, a pure terminal-based 3D renderer that displays Gaussian splats using Unicode and ASCII symbols, built over 3 days using 70-80 AI agents coordinated through a Claude Code setup with subagents inside subagents.

OpenClawRadar
Hypura: Storage-tier-aware LLM inference scheduler for Apple Silicon
Tools

Hypura: Storage-tier-aware LLM inference scheduler for Apple Silicon

Hypura is a Rust-based inference scheduler that places model tensors across GPU, RAM, and NVMe tiers to run models exceeding physical memory on Apple Silicon Macs. It enables running a 31GB Mixtral 8x7B on a 32GB Mac Mini at 2.2 tok/s and a 40GB Llama 70B at 0.3 tok/s where vanilla llama.cpp crashes.

OpenClawRadar
Team Brain: A Shared Memory Plugin for Claude Code That Stores Team Knowledge in Git
Tools

Team Brain: A Shared Memory Plugin for Claude Code That Stores Team Knowledge in Git

Team Brain is a Claude Code plugin that stores team knowledge in a .team-brain/ folder within your repository. It automatically generates a BRAIN.md file capped at 180 lines for optimal Claude instruction accuracy and works across tools by creating .cursorrules and AGENTS.md files.

OpenClawRadar
Super Claude browser extension makes Claude.ai UI fully customizable
Tools

Super Claude browser extension makes Claude.ai UI fully customizable

A developer built a browser extension that lets users customize every aspect of Claude.ai's interface — colors, fonts, layout, plus usage tracking and token counting. The extension works on Chrome and Firefox and was developed using Claude itself.

OpenClawRadar