Help us improve
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Share bugs, ideas, or general feedback.
Discover, evaluate (quality, licensing, provenance), and acquire datasets from Kaggle, HuggingFace, IPFS, arXiv, DBLP for AI model training and fine-tuning. Download free or paid data via local stdio MCP server—no API keys needed.
npx claudepluginhub guixu-project/guixuGuixu is the self-evolving AI agent for autonomous data discovery and trading. Guixu empowers AI agents with task-aware data discovery and decentralized data trading on Blockchain. AI agents can autonomously discover both free and paid data sources through a unified interface on Guixu.
curl -fsSL https://raw.githubusercontent.com/guixu-project/guixu/main/install.sh | bash
That's it. The installer downloads the binary, initializes a local node, auto-detects your AI clients (Codex, Cursor, Claude Code, OpenCode, OpenClaw), and registers the Guixu MCP server with each one.
After cloning the repository, run the bootstrap script to configure git hooks:
./bootstrap.sh
This configures core.hooksPath to use the project's .githooks/ directory, enabling automatic pre-commit checks (SPDX headers, English comments, formatting, clippy) before each commit.
Guixu exposes an MCP server so AI agents can discover, evaluate, and purchase datasets.
guixu mcp install codex # Codex
guixu mcp install cursor # Cursor
guixu mcp install claude-code # Claude Code
guixu mcp install opencode # OpenCode
guixu mcp install openclaw # OpenClaw
To remove: guixu mcp uninstall <client>
openclaw uses ~/.openclaw/config.json for MCP registration and installs a Guixu skill at ~/.openclaw/workspace/skills/guixu/.
Current OpenClaw support is installation-layer integration: Guixu registers a standard MCP server entry and installs a skill that nudges dataset acquisition requests through the Guixu workflow. The MCP tools themselves are shared across all clients.
Add to your client's MCP config:
{
"mcpServers": {
"guixu": {
"command": "guixu",
"args": ["mcp"]
}
}
}
| Tool | Description |
|---|---|
intent_parse | Parse natural-language data requests |
dataset_search | Search the Guixu network for datasets |
dataset_evaluate | Evaluate dataset suitability and value |
dataset_purchase | Purchase a dataset via smart contract |
dataset_feedback | Record post-use feedback on-chain |
guixu mcp --mode http # MCP over HTTP on :3927/rpc
Guixu can be packaged as a plugin for Codex Desktop and Claude Cowork.
Use the automated packaging script to create plugin packages:
# Package plugins for both Codex Desktop and Claude Cowork
make package
# Or use the script directly
./scripts/package-plugins.sh --output dist --clean
# Package only for Codex Desktop
make package-codex
# Package only for Claude Cowork
make package-claude
After packaging, install the plugins:
# Install Codex Desktop plugin
codex plugin install dist/guixu-codex-plugin
# Install Claude Cowork plugin
claude plugin install dist/guixu-claude-plugin
The packaged plugins follow the standard structure:
.codex-plugin/plugin.json + skills/ + .mcp.json.claude-plugin/plugin.json + skills/ + .mcp.jsonFor more details on plugin development, see guixu-plugin.md.
For packaging and installation instructions, see plugin-packaging.md.
Guixu is licensed under the Apache License, Version 2.0. See LICENSE for the full text and NOTICE for repository attribution notices.
Trademark use is governed separately. Guixu, the Guixu logo, and related marks are not licensed under Apache-2.0 except for reasonable and customary use in describing the origin of the software and reproducing the NOTICE file. See TRADEMARKS.md.
Contribution terms are described in CONTRIBUTING.md and CLA.md. Unless you explicitly state otherwise in writing and the maintainers agree, any contribution intentionally submitted for inclusion in this project is submitted under Apache-2.0, subject to any separate signed contributor agreement that applies.
Share bugs, ideas, or general feedback.
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Write SQL, explore datasets, and generate insights faster. Build visualizations and dashboards, and turn raw data into clear stories for stakeholders.
Give your AI a memory — mine projects and conversations into a searchable palace. 19 MCP tools, auto-save hooks, and guided setup.
MCP server that saves 98% of your context window with session continuity. Sandboxed code execution in 11 languages, FTS5 knowledge base with BM25 ranking, and automatic state restore across compactions.
Open-source, local-first Claude Code plugin for token reduction, context compression, and cost optimization using hybrid RAG retrieval (BM25 + vector search), reranking, AST-aware chunking, and compact context packets.
Agent Skills for AI/ML tasks including dataset creation, model training, evaluation, and research paper publishing on Hugging Face Hub
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge.
Sign in to claim