Back to list
ExcaliClaw Launches on Product Hunt: Bridging Collaborative Hand-Drawn Diagrams and AI Agents via Model Context Protocol
Product LaunchProduct HuntModel Context ProtocolDeveloper Tools

ExcaliClaw Launches on Product Hunt: Bridging Collaborative Hand-Drawn Diagrams and AI Agents via Model Context Protocol

Independent developer Mounir Bella has officially launched ExcaliClaw on Product Hunt, introducing a collaborative whiteboard platform designed to unify human diagramming with AI agent capabilities. Powered by the open-source Excalidraw editor, ExcaliClaw addresses longstanding workflow hurdles where AI coding companions are limited to ASCII art or text-based Mermaid code. Through a dedicated Model Context Protocol (MCP) server, developers can directly link assistants like Claude, Cursor, and VS Code via OAuth to create, edit, and organize persistent visual architecture boards in real time. Featuring instant search, collections, multi-language availability, and robust version recovery, ExcaliClaw establishes a dedicated space where developers and automated agents seamlessly co-author visual systems.

Product Hunt

Key Takeaways

  • Agent-Ready Visual Workspace: ExcaliClaw bridges the gap between text-based AI development and visual system architecture by allowing AI assistants to co-create and edit diagrams directly on an Excalidraw-powered canvas.
  • Model Context Protocol (MCP) Integration: Built-in MCP server connectivity enables seamless OAuth-based pairing with leading developer environments and AI assistants, including Anthropic's Claude, Cursor, and VS Code.
  • Persistent Organization & Collaboration: Resolves messy local file handling by offering centralized workspaces, instant search across diagrams, collections, link-based sharing, and real-time multi-user editing.
  • Accessible Pricing Tier Structure: Features a no-credit-card-required free tier offering up to 200 drawings and unlimited AI MCP usage, alongside Starter (€9/month) and Pro (€29/month) tiers.
  • Safety and Localization Built-In: Provides comprehensive version history, trash management, and backup export alongside native interface support for English, Spanish, and French.

In-Depth Analysis

Bridging AI Coding Assistants and Visual System Design

For software engineers and system architects, visual communication remains an essential component of design reviews and technical planning. However, the rise of large language model coding agents introduced an unexpected friction point: while AI assistants excel at generating source code and documentation, their diagramming capabilities have historically been restricted to static text representations like ASCII boxes or Mermaid script blocks. These text approximations frequently lack spatial intuitiveness and force developers to manually paste outputs into third-party diagramming tools.

Created by maker Mounir Bella, ExcaliClaw directly addresses this disconnect by providing an environment where visual diagrams are first-class citizens for both humans and AI models. By leveraging the familiar, intuitive aesthetics of the Excalidraw drawing canvas, ExcaliClaw allows autonomous agents to visualize complex architectures directly on the board. Developers no longer need to translate mental models back and forth between raw syntax and static rendering engines; instead, the AI agent interacts with the visual canvas as an active design participant.

Seamless Architecture via Model Context Protocol (MCP)

At the core of ExcaliClaw’s technical differentiation is its adoption of Anthropic's Model Context Protocol (MCP). The MCP implementation acts as an interoperable bridge between local developer environments and the cloud-hosted visual canvas. Through an authentication flow powered by OAuth, users can connect their ExcaliClaw workspace to AI-driven tools such as Claude, Cursor, or Visual Studio Code without manually copying API tokens or managing fragile configuration files.

Once connected, the AI agent gains programmatic tools to inspect, generate, and modify diagram elements within collections. This setup unlocks bidirectional collaboration: an engineer can sketch the broad boundaries of a microservices topology, highlight a specific service box, and prompt their connected IDE agent to break down the authentication subsystem inside the canvas. Because edits occur within an active diagram workspace rather than an isolated terminal, users retain full control to rearrange items, adjust hand-drawn styles, and refine details.

Organizational Infrastructure and Fault Tolerance

Beyond AI generation, ExcaliClaw tackles common collaboration and file-management challenges associated with ad-hoc whiteboarding. Standalone diagram files often end up scattered across repositories, local storage folders, or lost browser sessions. ExcaliClaw establishes a persistent workspace where diagrams are categorized into searchable collections, discoverable instantly by name, and shareable either via secure public links or team member usernames.

To ensure that autonomous AI modifications do not disrupt user workflows or destroy carefully arranged layouts, ExcaliClaw integrates continuous version history, trash recovery, and full backup export options. These safety nets make experimental edits virtually risk-free—if an AI agent overhauls a layout unexpectedly, users can easily roll back to previous revisions. Furthermore, broad accessibility is prioritized from day one, offering native user interfaces in English, Spanish, and French, alongside a zero-friction trial board that allows immediate sketch generation without account registration.

Industry Impact

ExcaliClaw’s launch marks an important milestone in the evolution of human-AI collaboration tools. As autonomous developer agents gain broader adoption across software engineering organizations, the primary interaction paradigm is shifting rapidly from one-dimensional text prompts toward rich, multi-modal collaboration canvases.

By leveraging the Model Context Protocol, ExcaliClaw highlights the growing significance of standardized open protocols over proprietary, closed plugin ecosystems. Instead of forcing developers into a siloed AI whiteboard product with walled-garden models, ExcaliClaw meets developers where they already work—inside environments like VS Code, Cursor, and Claude. This composable architecture demonstrates how dedicated software products can transform their user interfaces into API-accessible workspaces, setting a clear precedent for how productivity software, design tools, and IDEs will integrate autonomous agents in the future.

Frequently Asked Questions

What is ExcaliClaw and how does it integrate with AI assistants?

ExcaliClaw is an independent collaborative workspace built around the Excalidraw editor that provides persistent storage, real-time sharing, and diagram collections. It integrates with AI assistants via a hosted Model Context Protocol (MCP) server, allowing tools like Claude, Cursor, and VS Code to programmatically create and modify hand-drawn architecture diagrams through seamless OAuth authentication.

Which developer environments and languages are supported by ExcaliClaw?

ExcaliClaw supports primary MCP-compatible developer tools, notably Cursor, Anthropic's Claude, and Visual Studio Code. From a localization perspective, the user interface natively supports English, Spanish (Español), and French (Français).

What are ExcaliClaw's pricing tiers and AI usage allowances?

ExcaliClaw provides a generous Free tier that includes up to 200 saved drawings and unlimited AI MCP usage without requiring a credit card. For larger needs, the Starter tier is priced at €9 per month (unlimited drawings and 30-day version history), and the Pro tier is offered at €29 per month (unlimited history, site embedding capabilities, and extra team seats for €3 per month).

Related News

Gemini API Tool Retrieval Preview Gives Agents Access to Hundreds of Tools With Fewer Tokens
Product Launch

Gemini API Tool Retrieval Preview Gives Agents Access to Hundreds of Tools With Fewer Tokens

Google AI Blog announced that Tool Retrieval is now in preview in the Gemini API. The feature is presented as a way to give agents access to hundreds of tools while using fewer tokens. The announcement pairs those two points: the breadth of tool access and lower token use. It does not explain how Tool Retrieval works, specify how token use was measured, or quantify the reduction. It also does not provide an exact tool count beyond describing access to hundreds of tools. The stated release status is preview, and the announcement does not give a timeline for general availability.

LangSmith Tool Gateway Enters Public Beta for Secure, Governed AI Agent Tool Use
Product Launch

LangSmith Tool Gateway Enters Public Beta for Secure, Governed AI Agent Tool Use

LangChain announced that LangSmith Tool Gateway is in Public Beta. The gateway handles credentials, identity, and access for every agent tool call, with policies, approvals, and traces. The announcement frames secure, governed tool use around the individual call: its focus is on how agents access tools, rather than on what those tools or agents can do. The update identifies what the gateway covers but does not explain how credentials are managed, how access decisions are made, when approvals apply, or what information traces contain. The Public Beta is the only release status stated in the announcement.

Google’s Gemini Decision 1 Picks Answers Instead of Writing Them Across Five Input Types
Product Launch

Google’s Gemini Decision 1 Picks Answers Instead of Writing Them Across Five Input Types

Google’s Gemini Decision 1 is a model designed to pick answers rather than write them, according to Shittu Olumide’s report for KDnuggets. It reads text, images, video, audio, and PDFs in a single forward pass and supports up to 262,144 input tokens per request. The central distinction is between what the model reads and what it does with that input: it accepts several kinds of material, but its stated task is answer selection, not answer writing. The token figure describes the maximum input per request, not how much the model outputs. The report does not explain how answers are presented for selection or provide results showing how well the model performs.