Real-time mobile observability with unsampled on-device logs, session replay and AI agents that triage and investigate app issues autonomously.
Berd
Summary
Block's open-source desktop workspace for AI agents — one native app for Goose, Claude Code, Codex, Copilot, Amp and Cursor.
Screenshots
Description
Berd is an open-source desktop application from Block that puts every AI coding agent you use behind one native interface. It was built for Block's own staff, who were running Goose, Claude Code and Codex through separate apps and separate configuration files, and was released publicly under the Apache 2.0 licence in August 2026.
One workspace, any provider
On first launch Berd asks which provider should power it — Goose, Claude Code, Codex, GitHub Copilot, Amp or Cursor — and can install the one you pick. From there the sidebar is organised around work rather than around chat: Agents, Skills, Automations, Projects and Chats each get their own space, and the onboarding tailors a starter set of agents to the kind of work you do (engineering, legal, marketing, product, design or writing). The tagline the team uses is "less chatting, more building".
Portable skills and automations
Berd supports portable Agent Skills, so a skill folder written for one harness travels with you rather than being locked to a single vendor's app. Automations cover recurring work, and an experimental-features system gates the UI and workflow behaviour that is not yet stable. Enterprise deployments can ship with managed provider settings and an optional companion CLI.
Under the hood
It is a genuinely native application built on Tauri 2 and React 19 — not an Electron wrapper — and it speaks to the upstream Goose backend over an ACP WebSocket. Signed builds are published for macOS on Apple silicon and Linux on x86_64; Windows installers exist but currently ship without platform-native code signatures.
Worth knowing before you dive in
Berd is free and the source is public, but it is open source in the read-and-fork sense: the project explicitly does not accept external pull requests, and asks contributors to file well-formed issues instead. It is also young — the first public release was v0.6.0 — so expect the pace of change that implies.
Reviews
Similar App Suggestions
Autonomous QA agents that generate, run and self-heal Playwright end-to-end and API tests on every pull request, as code you own.
Cursor's Git forge for the agentic era — host repos, review pull requests and let coding agents work in the same place.
AI review agents that QA a website against your own checklist, then route the findings for client approval.
Turns backlog tickets into reviewed pull requests — background triage, a task board wired to live agent sessions, and MCP control from your own terminal.