Corporate blog post from Block (block.xyz/inside), unsigned — the displayed author is "Block" —, published on August 18, 2026, ~930 words, announcing the open-sourcing of Berd, Block's internal desktop application for working with agents, and laying out the design thesis that guided it: giving agents character "not only through roles, instructions, skills, and tools, but through distinctive visual identities" — hence the in-house animated characters, the "Gloopies".
By **Aucun auteur nommé** : le billet est signé **« Block »** — le champ *Author* de la page porte le nom de l'entreprise. Publié le **18 août 2026** sur `block.xyz/inside`// Source block.xyz ↗/Reading 2 min/.md// Auto-verified translation
Corporate blog post from Block (block.xyz/inside), unsigned, published on August 18, 2026, announcing the open-sourcing of Berd and laying out the design thesis that guided it.
What Berd is."Berd is a desktop application our teams use to work with AI agents across projects, skills, tools, and models." Born of an internal problem: Block had access to capable agents — goose, Claude Code, Codex — but each imposed "different interfaces, configuration systems, and ways of managing context". The conclusion drawn: "we didn't need another model or agent harness, we needed a consistent environment around them". Berd brings together conversations, files, folders, instructions, agents and skills around persistent projects, to stop rebuilding context for every task.
The design thesis. Give agents character — not only through roles, instructions, skills and tools, but through distinct visual identities, including a collection of animated characters, the "Gloopies". The problem invoked is that of the empty prompt box: "the product gives people little sense of how the agent is configured, which context and tools are available to it, and how it differs from another agent". The post places the approach in the lineage of Square and Cash App — bringing design where the category had none. But the problem stated is a configuration-legibility problem, and the avatar solves distinguishability; the text acknowledges this in one line it does not develop: "The avatars make the agent recognizable. Its role, skills, and tools make it useful."
The architecture. Berd descends from goose, the open source agent framework launched by Block in January 2025, contributed to the Agentic AI Foundation (Linux Foundation, December 2025) alongside MCP and AGENTS.md. Explicit division: "goose remains the open agent framework and runtime. Berd is a desktop application built around it. Berd connects to goose through the Agent Client Protocol." goose holds the agent loop, Berd holds the experience.
The sequel is Buzz. Berd served to explore solo work; "But work rarely stays private". What Berd showed — "private space, durable context, recognizable agent identities, reusable skills, visible configuration" — will feed Buzz, the shared human+agent space. "Start alone, then go multiplayer."
Caveats.No figures, no user testing, no license named; one isolated overreach ("create custom agents to do any task they want"); and a post whose title announces a retrospective while keeping the product in the present tense — Berd is not declared deprecated, but the roadmap points to Buzz.
Key takeaways
Date / source.August 18, 2026, corporate blog of Block (block.xyz/inside), unsigned post, ~930 words. Announces the open-sourcing of Berd (github.com/block/berd), no license named.
Key framing. the problem stated is a configuration-legibility problem — "how the agent is configured, which context and tools are available to it, and how it differs from another agent" — and the answer put forward is a visual identity (animated avatars, the "Gloopies"). The post itself distinguishes the two registers: "The avatars make the agent recognizable. Its role, skills, and tools make it useful." An avatar makes an agent distinguishable, not legible. ### The three-tier articulation Verbatim: "goose remains the open agent framework and runtime. Berd is a desktop application built around it. Berd connects to goose through the Agent Client Protocol. This separation lets Berd focus on the desktop experience, including projects, context, sessions, agents, and configuration, while goose handles the underlying agent loop." | Tier | Role | Boundary | |-------|------|-----------| | goose | agent framework and runtime, holds the agent loop | — | | Berd | desktop client: projects, context, sessions, agents, configuration | Agent Client Protocol | | Buzz | shared space where solo work becomes collaborative | not specified in the post |
Production occurrence of the Agent Client Protocol. — Zed's ACP, not to be confused with the Agentic Commerce Protocol: see [[girard-acp-deux-protocoles-un-sigle-2026-08-02]] for the disambiguation and [[agentclientprotocol-introduction-2026-08-02]] for the specification.
The separation thesis."we didn't need another model or agent harness, we needed a consistent environment around them" — the missing layer is neither the model nor the harness, it is the client. The post says nothing about how Berd connects to Buzz, nor whether the same ACP boundary is used there. ### The six requirements handed off to Buzz Verbatim: "Berd showed us the importance of private space, durable context, recognizable agent identities, reusable skills, visible configuration, and clearer visibility into an agent's configured context, tools, and capabilities. Buzz gives those ideas somewhere to go when individual work becomes collaborative." | # | Requirement | Demonstrated in the post? | |---|----------|----------------------------| | 1 | private space | no | | 2 | durable context | yes — persistent projects | | 3 | recognizable agent identities | yes — it is the subject of the text | | 4 | reusable skills | no | | 5 | visible configuration | no | | 6 | visibility into context, tools, capabilities | no | List usable as-is as an evaluation grid for an agent client. On skill portability across clients, cross-reference with [[janakiram-agent-platform-portability-contract-2026-07-20]] and [[google-agent-plugins-packaging-skills-mcp-2026-08-06]]. ### Affiliations and ecosystem
goose. is hosted at github.com/aaif-goose/goose — Agentic AI Foundation (aaif.io), a foundation under the aegis of the Linux Foundation, neutral governance.
The post refers to MCP and AGENTS.md (agents.md) as shared standards. ### Product thesis "Work with an agent often begins alone. You research, experiment, gather context, and shape an idea before it is ready for other people… But work rarely stays private." → "Start alone, then go multiplayer." This is the most falsifiable claim in the post, and the one that motivates the move from Berd to Buzz. ### Caveats
No figures. anywhere in the text: no user count, no repo count, no duration, no cost.
No license named. for an open-source announcement.
One overreaching sentence, without example or screenshot: "Just by chatting with Berd, users can easily create custom agents to do any task they want, even helping with everyday tasks like planning travel or shopping."
Attributed claims
"The avatars make the agent recognizable. Its role, skills, and tools make it useful."
— Block
"we didn't need another model or agent harness, we needed a consistent environment around them"
— Block
"Work with an agent often begins alone. But work rarely stays private."
— Block
"an agent's character is both visual and functional: agents can look distinct because they are"
— Block
The knowledge graph extracted from this fiche — 23 entities, 29 relations.
In this graph :Block · Berd · goose · Agent Client Protocol · Agentic AI Foundation · Buzz · identité visuelle d'agent · Gloopies · lisibilité de la configuration d'agent · distinguabilité des agents · boîte de saisie vide · client de bureau pour agents · séparation client / boucle d'agent · fragmentation des interfaces d'agents · reconstruction du contexte à chaque tâche · travail solo avec agents · travail collaboratif humains-agents · design comme différenciateur · Square · Cash App · Claude Code · Codex · confiance vérifiable dans un agent