ScienceClaw is a personal research assistant built with LangChain DeepAgents and AIO Sandbox infrastructure, adopting a completely new architecture beyond OpenClaw. It offers stronger security, better transparency, and a more user-friendly experience.
ScienceClaw is a personal research assistant built with LangChain DeepAgents and AIO Sandbox infrastructure, adopting a completely new architecture beyond OpenClaw. It offers stronger security, better transparency, and a more user-friendly experience.
First-time install walkthrough for Claude Code, Codex CLI, and ChatGPT.
β¨ Why ScienceClaw
π Security First
ScienceClaw runs entirely inside Docker containers. The agent cannot access your host system, personal files, or environment variables. All code execution happens in an isolated sandbox, and generated data stays in a local ./workspace directory β nothing is uploaded to external servers. Deploy with confidence on your own machine.
ποΈ Full Transparency
Every step of the agent's workflow is visible and traceable β from web search and data crawling, to reasoning and tool invocation, to final report generation. You always know where results come from, what actions were taken, and how conclusions were reached. Inspect any step at any time.
π Ready Out of the Box
No tedious configuration needed. ScienceClaw ships with curated tools and skill packages β launch the entire environment with a single command. Whether you're a researcher, developer, or student, you can get started immediately. Focus on real tasks, not setup.
ποΈ Architecture
π’ News
[2026-03-13] ScienceClaw v0.0.1 is officially released! Visit our website: scienceclaw.taichuai.cn
π¦ Quick Start
Windows Users β Desktop App (One-Click Install)
No Docker required, no command-line needed. Download the desktop installer and get started instantly.
ScienceClaw makes it easy to extend with your own tools:
Natural language creation β Simply describe what you want in chat, and the agent will create, test, and save a new tool for you automatically.
Manual mounting β Drop any Python file with @tool decorated functions into the Tools/ directory; they are auto-detected and hot-loaded without restart.
π§ Skill System
Skills are structured instruction documents (SKILL.md) that guide the Agent through complex, multi-step workflows. Unlike tools (executable code), skills act as the Agent's "playbook" β defining strategies, rules, and best practices.
Built-in Skills
| Skill | Purpose |
|---|---|
| π pdf | Read, create, merge, split, OCR, and generate professional PDF research reports |
| π docx | Create and edit Word documents with cover pages, TOC, tables, and charts |
| π pptx | Generate and edit PowerPoint slide decks |
| π xlsx | Create and manipulate Excel spreadsheets, CSV/TSV data processing |
| π οΈ tool-creator | Create & upgrade custom @tool functions (write β test β save) |
| π skill-creator | Create & refine skills with draft β test β review β iterate |
| π find-skills | Search, discover & install community skills from the open ecosystem |
| π§ͺ tooluniverse | Unified access to 1,900+ scientific tools |
Multi-Format Report Generation
ScienceClaw can produce professional research deliverables in 4 document formats:
| Format | Features |
|---|---|
| PDF | Cover page, table of contents, charts (bar/pie/line), in-text citations, references, academic styling |
| DOCX | Cover page, TOC, tables, images,
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
