Editions: Community vs Hosted
Scrapalot follows an open-core model. The core is free and open source so you can self-host it on your own hardware, while the full-featured product is offered as a managed subscription on our cloud — the hosting, the advanced AI surfaces, and the native apps are what fund ongoing development.
There are two ways to run Scrapalot:
| Community Edition | Hosted (Pro · Team · Enterprise) | |
|---|---|---|
| Price | Free (self-hosted, you pay your own infra) | Free Researcher tier, paid Pro · Team · Enterprise — see Pricing |
| License | AGPL-3.0 (open source) | Proprietary, managed by us |
| Runs on | Your own server (Docker Compose) | Our cloud, zero setup |
| Apps | Browser | Browser · desktop (Windows/macOS/Linux) · Android |
| Updates & backups | You manage | Managed, automatic |
| Support | Community (Discussions, Discord) | Email / priority / dedicated |
The Community Edition source lives in its own public repository (AGPL-3.0): github.com/sime2408/scrapalot — all four services plus a
docker-compose.ymlthat runs the whole stack. The full hosted product is a separate, proprietary codebase. (⭐ star the repo and join Discord to follow development.)
What's in the free, open-source Community Edition
Everything you need to read, organize and chat with your own document library on your own infrastructure:
- 📥 Document ingestion — PDF and common formats, OCR, chunking and embeddings (pgvector)
- 🔎 Core RAG chat — vector-search retrieval over your collections
- 📝 Collaborative notes editor — TipTap with real-time (Y.js) editing
- 🔊 Read-aloud (TTS)
- 🔑 Bring Your Own Key (BYOK) — OpenAI, Anthropic, DeepSeek, or local models
- 🐳 Self-host with Docker Compose — single-user, no usage quotas on your own box
What's exclusive to the Hosted product
The advanced AI surfaces and collaboration features that make up the paid tiers stay in the proprietary, hosted product:
- 🧠 Advanced RAG — fusion, agentic routing and reranking strategies beyond core vector search
- 🔬 Deep Research — multi-phase, multi-agent, cited research runs
- 🕸️ Knowledge Graph — Neo4j entity linking and graph-aware retrieval
- 📄 AI Scientist — publication-style paper generation
- ✍️ Notes AI assistant — fact-check, peer review, compose from sources
- 🎙️ Voice chat / speech-to-text and 🖼️ image generation
- 🔌 MCP integrations (connect your own MCP servers, or use Scrapalot from Claude Code / Gemini CLI) and external connectors (Google Drive, Dropbox, OneDrive, Notion, Confluence, Slack, Zotero)
- 👥 Team chat, direct messages, huddles, shared workspaces and multi-user collaboration
- 🧮 Programmatic API access and advanced analytics
- 📱 Native apps — desktop (Windows, macOS, Linux, with local speech-to-text and local models) and Android
- 🛟 SLA, on-premise and enterprise options
Want it managed?
Skip the setup — start free on our cloud with the Researcher tier and upgrade when you need the advanced features. See Pricing.