NeuralScio documentation covers the native desktop workflow: scholarly search, saved papers, PDF reading, citation graph exploration, AI assistance, Neural Writer, and MCP/headless automation.
Live website docs are sourced from apps/web/src/docs/*.md. docsCatalog.ts controls order, titles, and related links. The apps/web/docs directory is internal web engineering documentation, not the live website content source.
Start Here
If you are new, read in this order:
1. Getting Started for download, installation, and the first desktop workflow. 2. Interface Tour to learn the main research surfaces. 3. Local AI Setup to choose the built-in GGUF worker or the separate Ollama-compatible path. 4. AI Features for BYOK, local AI, and the Public Alpha cloud boundary. 5. Offline Mode for the local-first boundary. 6. Troubleshooting or Support if you need help.
Other guides:
- MCP & Headless
- Agents
- Agent Research
- Cluster Map
- Neural Writer
- Scholarly Search
- Data Providers
- PDF Viewer
- Citation Graph
- Offline Mode
- Sync & Cloud
- Workspace Customization
- Settings
- Export Citations
- Voice Interface
- Keyboard Shortcuts
Product Surfaces That Matter Most
Native desktop foundation
- Native desktop application built in C++ and Qt
- Local settings, local caches, and a separate AI worker process in the desktop architecture
- Desktop UI plus non-UI launch paths in the same product
- Customizable themes, local PNG backgrounds, dock/tab visibility controls, and desktop status indicators where available
- Windows and Linux are the clearest active build paths; macOS support is more limited
Research workflow
- Scholarly search across 10+ sources with different access profiles
- Interface Tour article for mapping visible buttons, tabs, cards, docks, and setup actions
- Built-in PDF reading with annotations, search inside open PDFs, and cached reading
- Saved-library workflow for articles and local PDFs
- Citation graph exploration with 2D and 3D views
- Neural Writer for LaTeX drafting, bibliography sync, preview, and citation-aware writing
AI and automation
- Cloud AI providers: OpenAI, Anthropic Claude, Google Gemini, DeepSeek, Kimi, GLM, Qwen, and MiniMax
- Local AI paths: the built-in NeuralScioAI worker with GGUF models, plus the separate Ollama-compatible provider path
- MCP over stdio for local agent tooling
- MCP over HTTP and SSE for networked integrations
- Headless mode for scripted and agent-driven launches
Local vs Network Boundary
Strong local path
- Saved PDFs and local PDF library
- Local annotations and cached reading state
- Local AI workflows when you use the built-in worker or the Ollama-compatible provider
- Headless + stdio MCP workflows on the same machine
Still depends on network access
- Provider search across external scholarly sources
- Managed cloud AI
- Account-linked sync and billing flows
- Remote MCP connections when you choose the network transport
Access Modes
NeuralScio is in Public Alpha. The desktop app is free during alpha. Paid subscriptions are not active yet.
| Path | What it means during Public Alpha | | --- | --- | | Local Alpha | Free desktop access: saved/cached papers, your own documents, local workflows | | BYOK | Bring your own keys or local providers where supported; not a paid NeuralScio plan | | Pro | Planned after Public Beta; future managed cloud plan | | Team | Planned after Public Beta; future team/lab billing |
Local Alpha is not the same as managed online search or hosted cloud AI. See Pricing for details.
Where To Go Next
1. Read Getting Started and Interface Tour for the core desktop workflow. 2. Read Local AI Setup before configuring local models. 3. Read Pricing for Public Alpha access and future cloud/billing plans.