Technical Deep Dive
Excalidraw’s technical architecture is a masterclass in doing more with less. At its core, the application is a single-page React application that leverages a custom rendering engine built on top of the HTML5 Canvas API. The decision to use Canvas rather than SVG or DOM-based rendering was deliberate: Canvas provides superior performance for the kind of highly interactive, repaint-heavy operations that a whiteboard demands—zooming, panning, and dragging dozens of elements simultaneously. The team optimized rendering by implementing a dirty-rectangle approach, where only the changed portions of the canvas are redrawn, rather than the entire scene. This keeps frame rates high even on lower-end devices.
For state management, Excalidraw uses a centralized store built with zustand, a lightweight React state management library. Every action—adding a rectangle, moving a sticky note, changing a stroke color—dispatches a single action that updates the store immutably. This design makes undo/redo trivial (the app maintains a stack of previous states) and enables real-time collaboration to work seamlessly: each remote action is simply applied as a local state mutation via WebSocket or Firebase Realtime Database.
The hand-drawn aesthetic is achieved through a clever combination of cubic Bezier curves and random jitter. Rather than drawing perfect straight lines or rectangles, Excalidraw’s rendering function adds small, pseudo-random offsets to control points. The algorithm is deterministic per seed, so the same diagram always looks the same, but the overall effect is a charmingly imperfect, human-like sketch. This is not just cosmetic—studies in cognitive psychology suggest that hand-drawn diagrams are more memorable and less intimidating than polished ones, encouraging iteration and feedback.
Real-time collaboration is built on a Conflict-Free Replicated Data Type (CRDT) model, specifically using the Yjs library. Yjs allows multiple users to edit the same canvas simultaneously without a central server mediating every operation. Each user maintains a local copy of the document, and changes are merged using a last-write-wins strategy for conflicting edits. The system supports presence awareness (seeing other users’ cursors) and lazy loading of large canvases by only syncing the viewport area. The result is a collaboration experience that feels nearly as snappy as local editing, even with dozens of concurrent users.
For those who want to dig deeper, the entire codebase is available on GitHub under the MIT license. The repository is well-structured, with clear separation between the core library (`@excalidraw/excalidraw`) and the standalone application. The core library can be embedded in any React application, which is how tools like Obsidian, Notion, and various documentation platforms integrate Excalidraw. The project also maintains a growing collection of libraries and utilities, including an `excalidraw-cli` for programmatic diagram generation and a `@excalidraw/utils` package for parsing and validating `.excalidraw` files.
Data Takeaway: The combination of Canvas rendering, zustand state management, and Yjs CRDTs gives Excalidraw a performance profile that rivals native apps, while keeping the codebase maintainable and extensible. This technical foundation is the bedrock of its viral adoption.
Key Players & Case Studies
Excalidraw was created by Vjeux (Christopher Chedeau), a former Facebook engineer known for his work on React Native and Prettier. The project quickly attracted contributions from a wide community, with notable early adopters including the React core team, who used it for conference talks and documentation. Today, the project is maintained by a small group of volunteer maintainers, with occasional sponsorship from companies like Vercel and Netlify for hosting and CI/CD resources.
| Feature | Excalidraw | Miro | FigJam | tldraw |
|---|---|---|---|---|
| Pricing | Free (self-host or hosted) | $8/user/month (Starter) | Free (limited) / $3/user/month | Free (open source) |
| Open Source | Yes (MIT) | No | No | Yes (MIT) |
| Self-Hostable | Yes | No | No | Yes |
| Real-Time Collaboration | Yes (Yjs) | Yes | Yes | Yes (Yjs) |
| Hand-Drawn Style | Core feature | No | No | Optional (plugin) |
| Embeddable | Yes (React component) | Limited (iframe) | Limited (iframe) | Yes (React component) |
| GitHub Stars | 122,000+ | N/A | N/A | 30,000+ |
| Typical Use Case | Quick diagrams, architecture sketches | Brainstorming, workshops | Design sprints, wireframes | Whiteboarding, drawing |
Data Takeaway: Excalidraw’s main competitive advantage is its zero-cost, open-source nature combined with a unique hand-drawn aesthetic. While Miro and FigJam offer richer collaboration features (like voting, timers, and extensive template libraries), Excalidraw wins on simplicity, embeddability, and the trust that comes from being fully auditable. tldraw is its closest open-source competitor, but Excalidraw’s larger community and earlier market entry give it a network effect advantage.
Case Study: The Obsidian integration is a prime example of Excalidraw’s embeddability. Obsidian, a popular note-taking app, allows users to embed Excalidraw canvases directly into their notes. This has made Excalidraw the default diagramming tool for the Obsidian ecosystem, which has over 1 million monthly active users. Similarly, the Notion integration (via a third-party widget) has driven significant adoption among product teams who use Notion as their knowledge base.
Industry Impact & Market Dynamics
The virtual whiteboard market was valued at approximately $2.5 billion in 2024 and is projected to grow to over $8 billion by 2030, driven by the permanent shift to hybrid and remote work. Excalidraw occupies a unique niche: it is not a full-fledged collaboration platform like Miro, nor a design tool like Figma. Instead, it is a lightweight, embeddable component that serves as the “diagram engine” for other tools. This positioning allows it to avoid direct competition with the giants while benefiting from their growth.
| Metric | Excalidraw | Miro (2024 est.) | FigJam (2024 est.) |
|---|---|---|---|
| Monthly Active Users | 5-10M (estimated) | 70M+ | 30M+ |
| Revenue Model | Donations / Sponsorships | SaaS subscriptions | Part of Figma suite |
| Annual Revenue | <$1M (estimated) | $400M+ | Bundled with Figma |
| GitHub Stars Growth (YoY) | ~40% | N/A | N/A |
| Enterprise Adoption | Low (no SSO, no admin) | High | Medium |
Data Takeaway: Excalidraw’s user base is large but monetization is virtually nonexistent. This is both a strength (no pressure to enshittify the product) and a vulnerability (no sustainable funding for full-time development). The project relies on the goodwill of maintainers and occasional corporate sponsorships. The risk is that if key maintainers burn out, the project could stagnate.
The broader market trend is toward composable tools. Companies are increasingly building custom internal tools by stitching together open-source components like Excalidraw, CodeMirror (for code editing), and TipTap (for rich text). This “Lego block” approach to software is challenging the all-in-one suite model. Excalidraw’s embeddable React component is a perfect example: it can be dropped into any web application with a single npm install command, giving any product instant diagramming capabilities.
Risks, Limitations & Open Questions
Despite its success, Excalidraw faces several critical challenges:
1. Sustainability: The project has no clear revenue stream. The maintainers have resisted adding paid features, which is admirable but risky. If the project cannot fund itself, it may eventually fall behind on security patches, performance improvements, or feature parity with competitors. The recent addition of a sponsorship button is a step, but it is unlikely to generate the six-figure sums needed for full-time development.
2. Feature Creep: The community regularly requests advanced features—layers, grouping, alignment guides, vector editing, shape libraries. Each new feature adds complexity and risks diluting the core simplicity that made Excalidraw popular. The maintainers have been disciplined so far, but the pressure is mounting.
3. Enterprise Gaps: Excalidraw lacks enterprise-grade features like single sign-on (SSO), audit logs, role-based access control, and compliance certifications (SOC 2, HIPAA). This limits its adoption in regulated industries and large organizations. Miro and Lucidchart have built their businesses on these features.
4. Competition from AI: AI-powered diagramming tools (e.g., diagrams.net with AI plugins, or Microsoft Visio’s AI features) can generate diagrams from natural language prompts. Excalidraw has no native AI integration, which could make it feel outdated as users expect intelligent assistance.
5. Security of Self-Hosted Instances: While self-hosting is a feature, it also means that users are responsible for their own security. The project provides no official guidance on hardening deployments, and many self-hosted instances run on default configurations, potentially exposing sensitive data.
AINews Verdict & Predictions
Excalidraw is a rare gem in the open-source ecosystem: a tool that is simultaneously simple, powerful, and beloved. Its success proves that there is still room for minimalist, well-designed software in an industry obsessed with feature bloat. However, the project is at a crossroads. The next 12-18 months will determine whether it remains a hobbyist darling or evolves into a sustainable, professionally maintained project.
Predictions:
1. Excalidraw will be acquired by a larger company within 2 years. The most likely acquirers are Notion (to deepen its diagramming capabilities), GitHub (to integrate into Issues and Discussions), or a cloud provider like Vercel or Netlify (to offer a hosted version as part of their platform). The acquisition price would likely be in the $10-30 million range, based on comparable open-source acquisitions.
2. The project will add a lightweight AI feature by mid-2026. This will likely be a “generate diagram from text” feature using a small, locally-run model (e.g., Llama 3.2 or Gemma) to preserve the offline-first ethos. It will be opt-in and not require an API key.
3. A hosted, paid tier will emerge, but it will be optional. The hosted version will offer SSO, team management, and priority support, while the self-hosted version remains free and fully featured. This is the model used by GitLab and Mattermost, and it strikes the right balance between openness and sustainability.
4. Excalidraw will become the default diagramming engine for AI coding assistants. Tools like Cursor, GitHub Copilot, and Claude Artifacts will integrate Excalidraw to let users visualize code architecture and data flows. This will drive a new wave of adoption among developers.
What to watch: The health of the GitHub repository. If the maintainers start rejecting reasonable PRs or the issue backlog grows beyond 500 open issues, it may signal that the project is under-resourced. Conversely, if a major company announces a sponsorship or integration, it will validate the project’s long-term viability. For now, Excalidraw remains the best tool for anyone who needs to sketch an idea quickly and share it without friction—and that is a remarkable achievement for a project built on goodwill and good code.