Rosie VS SpawnGraph

SpawnGraph's answer:

SpawnGraph is the only mind mapping tool that combines three things in one product: (1) browser-native natural-language processing — every PDF, document, and transcript is parsed entirely client-side, so your content never reaches our servers and you can verify it in DevTools → Network; (2) automatic mind map generation from 93+ source file formats including PDFs, Word, PowerPoint, EPUBs, YouTube transcripts (auto-fetched), audio recordings (Whisper WASM), and images (Tesseract OCR); and (3) real-time multi-user collaboration with named cursors on every plan including the free tier, with no per-seat pricing. Every other tool in the category sends your content to a cloud AI service, charges per generation, or locks collaboration behind a paid tier.

SpawnGraph's answer:

Three reasons:

1. Speed-to-first-map. Drop a 30-page research paper and SpawnGraph produces a hierarchical mind map in under 2 seconds. Tools like MindMeister, Miro, and Whimsical require you to type every node yourself. AI tools like Mapify or MyMap upload your document, charge per generation, and wait several seconds for a model round-trip.

2. Genuine privacy. Confidential pre-prints, interview transcripts, legal contracts, embargoed materials — none of it leaves your browser during generation. This isn't a marketing claim; open DevTools → Network and watch zero outbound requests during a generation. Critical for researchers, lawyers, and anyone working with sensitive content.

3. Free tier is the product, not a trial. Miro caps you at 3 editable boards. MindMeister at 3 maps. Heptabase has no free tier at all — just a 7-day trial. SpawnGraph's free tier covers 2 saved mindmaps, 8 export formats including watermarked PNG, real-time collaboration with up to 5 editors per board, and unlimited generation from any of 93+ file formats. Solo Pro at $5/month annual unlocks unlimited boards and the full export set — but the free tier is genuinely usable forever.

SpawnGraph's answer:

SpawnGraph's primary users are knowledge workers who deal with a lot of source material and need to extract structure from it quickly:

• Students reviewing lecture PDFs, textbook chapters, and YouTube lectures during exam season
• Researchers and PhD students conducting literature reviews where each paper needs to be understood structurally before it can be synthesized into a thesis
• Consultants turning client transcripts, briefing documents, and research reports into shareable visual frameworks
• Content creators mapping a long-form video, podcast, or article before producing derivative content
• Product managers and engineers mapping the structure of technical specs, RFCs, and meeting notes

Secondary audience: PKM enthusiasts (Obsidian, Heptabase, Logseq users) looking for a visual synthesis tool that complements their long-term knowledge management vault.

SpawnGraph's answer:

I built SpawnGraph because every "AI mind map" tool I tried in 2024–2025 had three problems: they required signup, they uploaded my PDFs to their servers, and they charged per generation (sometimes hidden as "credits"). I wanted to map research papers, technical docs, and YouTube talks without any of those constraints.

The technical insight was that modern browsers can do most of the heavy lifting client-side: pdf.js for PDF text extraction, Intl.Segmenter for tokenization (built into V8), heading-hierarchy reconstruction in plain JavaScript, Whisper.cpp WASM for audio transcription, and Tesseract.js for image OCR. The "AI" part of AI mind maps was largely unnecessary — most real-world documents already have implicit structure (headings, paragraphs, lists) that can be extracted deterministically.

Real-time collaboration was added next because mind mapping is genuinely better as a multi-player activity. We used Yjs CRDTs over a custom binary WebSocket protocol — the same underlying tech as Linear, Figma's multiplayer layer, and AFFiNE.

The product is solo-built and bootstrapped. We're currently pivoting toward a local-first desktop model (Tauri 2.x) for power users who want the same workflow without any browser dependency, modeled on Obsidian's approach.

SpawnGraph's answer:

Frontend (browser): - React 19 + Vite + TypeScript - @xyflow/react for the canvas - Yjs + y-indexeddb + y-protocols for real-time CRDT collaboration - Zustand for client state - Tailwind CSS for styling - pdf.js for PDF text extraction - Intl.Segmenter for tokenization - Whisper.cpp WASM for audio transcription - Tesseract.js for image OCR - @xenova/transformers for browser-side ML utilities - pdf-lib + pptxgenjs + jszip for client-side export to PDF/PPTX/zip - React Router v6 + SSR prerender for marketing pages

Backend: - Hono framework on Node.js (single VPS) - Postgres for primary storage - Lucia Auth for session management - Yjs sync server (WebSocket binary protocol) - Dodo Payments for billing (not Stripe) - Cloudflare Pages for static hosting + edge caching - Cloudflare DNS + Cloudflare Insights for analytics

Infra cost is ~$30/month at current traffic — single VPS + Cloudflare free tier + small Postgres instance.

SpawnGraph's answer:

• SpawnGraph is an early-stage product currently building its customer base. Users include independent researchers, university students, indie consultants, and small product teams. We'll update this section as we get permission to name specific accounts.