A slide framework built for agents.
5.7k
Stars
398
Forks
61
Open issues
15
Contributors
AI Analysis
open-slide is a TypeScript framework for building presentation decks optimized for AI agent authoring, where agents write React components that render to a fixed 1920×1080 canvas. It's purpose-built for the emerging workflow of natural-language-to-slides generation via Claude Code, Cursor, and similar coding agents—not a general presentation tool for end users. Best suited for developers and agent builders who want agent-native slide creation with built-in inspector, asset management, and pre...
Inferred from signals mentioned in the README (tests, CI, type safety) — not a review of the actual code.
AI's overall editorial judgment — not an average of the bars above, can weigh other factors too.
open-slide: A React-based slide runtime designed to be authored by AI coding agents
open-slide is a TypeScript/React slide framework that positions itself as the runtime layer for AI-driven presentation creation. Instead of a proprietary DSL, every slide is a plain React component rendered on a fixed 1920×1080 canvas. The target user is a developer or technically-inclined knowledge worker who wants to describe a deck in natural language to a coding agent (Claude Code, Cursor, Codex, etc.) and receive a polished, deployable result. It handles canvas scaling, hot reload, a presenter mode with speaker notes and timer, PDF/static HTML export, and an in-browser comment inspector that feeds back into the agent loop. It is not a competitor to Canva or PowerPoint — it is explicitly a developer tool that assumes agent-mediated authoring.
Created in late April 2026, open-slide is very young (~2 months old). It emerged alongside the broader wave of agent-native tooling that accelerated in 2025-2026, explicitly targeting the moment when coding agents became capable enough to generate polished UI code reliably.
The project reached 5,480 stars in roughly 8 weeks, suggesting a strong initial launch spike — likely driven by social media visibility in the developer and AI tooling communities. Growth has since settled to approximately 49 stars per week, indicating the initial viral moment has passed but interest remains active. Vercel OSS Program participation provides some institutional visibility.
Adoption not verified in terms of documented production deployments or case studies. The Vercel OSS Program badge implies some level of vetting by Vercel, but this is a low bar for adoption evidence. Fork count of 376 on a 2-month-old repo suggests active experimentation, but real-world production use at scale has not been publicly documented.
Appears to be a pnpm + Turborepo monorepo with two main packages: @open-slide/core (the runtime, Vite plugin, dev/build/preview CLI) and @open-slide/cli (the npx scaffolder). Slides are arbitrary React components, meaning no proprietary DSL — this is a deliberate architectural choice to remain agent-friendly. The core likely wraps a Vite dev server with custom middleware for the inspector. Static export is likely implemented via Vite's build mode. Agent skills appear to be CLAUDE.md-style prompt files bundled at scaffold time, not runtime code.
Not documented in README
Last push was 2026-06-21, two days before the evaluation date — actively maintained. The project is 8 weeks old and receiving frequent updates, which is expected for an early-stage project. No information about issue response times or contributor count beyond the fork count (376), which suggests moderate external engagement for a project this young.
ADOPT IF: you regularly use AI coding agents to create technical presentations and want a structured runtime that gives agents clear constraints (fixed canvas, type scale, layout rules) and a feedback loop via in-browser commenting. AVOID IF: you need a mature, community-supported ecosystem with plugins, themes, and proven enterprise use — at 8 weeks old, open-slide carries early-adopter risk including potential breaking changes, sparse documentation, and limited community support. MONITOR IF: you are interested in agent-native developer tooling broadly and want to track whether the agent-authoring workflow matures into something with verified production adoption over the next 6-12 months.
Independent dimensions
Mainstream potential
4/10
Technical importance
6/10
Adoption evidence
2/10
- Project is only 8 weeks old — APIs and conventions may change substantially before stabilizing, making early adoption a maintenance liability.
- The 'agent-native' positioning depends on the continued capability and availability of specific coding agents (Claude Code, Cursor, Codex); shifts in those products could reduce the value proposition.
- Single-maintainer risk appears plausible given the early stage — no evidence of a broader core team is visible in the README, though this cannot be confirmed from metadata alone.
- The fixed 1920×1080 canvas is a deliberate constraint but may limit use cases requiring responsive or adaptive presentations without significant workarounds.
- Adoption appears largely driven by initial launch interest; if the weekly growth rate continues declining, the project may struggle to build the community momentum needed to sustain long-term development.
Likely to develop a loyal niche among developer-heavy teams that live in coding agents. Mainstream adoption beyond that niche will depend on whether agent-mediated slide creation becomes a common workflow — currently an open question.
Newsletter
Get analyses like this every Monday
Free weekly digest of the most interesting open-source discoveries.
Languages
Information
- Website
- https://open-slide.dev
- Language
- TypeScript
- License
- MIT
- Last updated
- 3d ago
- Created
- 3mo ago
- Analyzed with
- anthropic/claude-haiku-4-5
Stars over time
Contributors over time
Top 100 contributors only — repos with more will plateau at 100.
Open issues
Top contributors
Similar repos
zarazhangrui/frontend-slides
Frontend Slides is a Claude Code plugin that generates beautiful HTML...
presenton/presenton
Presenton is a self-hosted, open-source AI presentation generator that creates...
openai/openai-agents-js
The OpenAI Agents SDK is a TypeScript framework for building multi-agent AI...
| Repository | Stars | Week Δ | Language | Score | Updated |
|---|---|---|---|---|---|
|
|
5.7k | +78 | TypeScript | 8/10 | 3d ago |
|
|
25k | — | JavaScript | 8/10 | 2w ago |
|
|
9k | — | TypeScript | 7/10 | 21h ago |
|
|
47.6k | — | TypeScript | 8/10 | 9h ago |
|
|
3.4k | — | TypeScript | 8/10 | 12h ago |
|
|
3.6k | — | HTML | 7/10 | 1mo ago |
Slidev is the dominant developer-centric slide tool with 47K+ stars and a large ecosystem. It uses Markdown with Vue components. open-slide targets a narrower use case: agent-mediated authoring with React. For human-driven authoring, Slidev's ecosystem and community are substantially larger. For agent authoring, open-slide's React-native approach may be more compatible with how agents generate code.
Presenton (8.4K stars) appears to target AI-assisted presentation creation with a different angle — likely a more GUI-oriented approach. open-slide is explicitly code-first and developer-targeted. The two may serve overlapping audiences but with different workflows.
With 22K+ stars, this is a large but likely content-specific repo (possibly a collection of slides rather than a framework). Not a direct competitor to open-slide as a runtime framework.
15K stars suggests a broader agent-to-UI category tool. open-slide is narrowly focused on presentations, while a2ui appears to target general UI generation. They solve adjacent but distinct problems.
This is an agent SDK, not a slide tool — it is infrastructure that open-slide's agent skills could theoretically be built on top of. Not a competitor; more of a potential integration point.