robot-tool-mcp
Provides an MCP interface to the ROBOT command-line tool for OWL ontology editing, enabling operations like merging, reasoning, and conversion via natural language.
README
ROBOT Ontology MCP
MCP server wrapping the ROBOT command-line tool for OWL ontology editing.
Prerequisites
ROBOT must be installed and available on your PATH:
robot --version
See ROBOT installation for setup instructions.
Installation
With uvx (recommended)
uvx robot-tool-mcp
From source
git clone https://github.com/<owner>/robot-mcp.git
cd robot-mcp
uv run robot-mcp
Claude Desktop Configuration
Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"robot-mcp": {
"command": "uvx",
"args": ["robot-tool-mcp"]
}
}
}
Or for local development:
{
"mcpServers": {
"robot-mcp": {
"command": "uv",
"args": ["run", "--directory", "/path/to/robot-tool-mcp", "robot-mcp"]
}
}
}
Tools
Individual Command Tools (25 tools)
Each ROBOT command is exposed as a dedicated MCP tool with typed parameters:
| Tool | Description |
|---|---|
robot_annotate |
Add metadata annotations to an ontology |
robot_collapse |
Simplify class hierarchies by removing intermediates |
robot_convert |
Transform ontology between formats (OWL, OBO, TTL, etc.) |
robot_diff |
Compare two ontologies semantically |
robot_expand |
Convert shortcut annotations into OWL axioms |
robot_explain |
Debug inferred statements with minimal explanations |
robot_export |
Generate tabular output (CSV, TSV, JSON, HTML, XLSX) |
robot_extract |
Create a subset module (STAR, BOT, TOP, MIREOT) |
robot_filter |
Selectively copy axioms (inverse of remove) |
robot_materialize |
Assert inferred superclass relationships |
robot_measure |
Compute ontology metrics and statistics |
robot_merge |
Combine multiple ontologies into one |
robot_mirror |
Cache imported ontologies locally |
robot_query |
Execute SPARQL queries (SELECT, ASK, CONSTRUCT, UPDATE) |
robot_reason |
Run OWL reasoner (ELK, HermiT, JFact, Whelk) |
robot_reduce |
Remove redundant subClassOf axioms |
robot_relax |
Convert equivalence axioms to subclass axioms |
robot_remove |
Eliminate selected axioms (inverse of filter) |
robot_rename |
Modify entity IRIs |
robot_repair |
Fix common ontology problems |
robot_report |
Run quality control checks with violation report |
robot_template |
Convert tabular data (CSV/TSV) into OWL |
robot_unmerge |
Remove axioms of one ontology from another |
robot_validate_profile |
Check OWL 2 profile conformance (EL/RL/QL/DL) |
robot_verify |
Check ontology against SPARQL rules |
Chain Tool
The robot_chain tool pipelines multiple commands in a single ROBOT process. Ontology objects pass in-memory between steps — no intermediate files needed.
{
"steps": [
{"command": "merge", "input": ["edit.owl", "base.owl"]},
{"command": "reason", "reasoner": "ELK"},
{"command": "annotate", "ontology_iri": "https://example.org/my.owl"},
{"command": "convert", "format": "ofn", "output": "result.owl"}
]
}
This produces a single CLI call:
robot merge --input edit.owl --input base.owl \
reason --reasoner ELK \
annotate --ontology-iri https://example.org/my.owl \
convert --format ofn --output result.owl
Argument mapping rules:
- Underscores become hyphens:
ontology_iri→--ontology-iri - Lists repeat the flag:
{"input": ["a.owl", "b.owl"]}→--input a.owl --input b.owl - Booleans become strings:
true→"true"
Common Workflows
Build a release:
{
"steps": [
{"command": "merge", "input": ["edit.owl"]},
{"command": "reason", "reasoner": "ELK"},
{"command": "relax"},
{"command": "reduce", "reasoner": "ELK"},
{"command": "annotate", "ontology_iri": "https://example.org/release.owl",
"version_iri": "https://example.org/2024-01-01/release.owl"},
{"command": "convert", "output": "release.owl"}
]
}
Extract a module:
{
"steps": [
{"command": "merge", "input": ["full-ontology.owl"]},
{"command": "extract", "method": "BOT", "term": ["GO:0005634", "GO:0005737"]},
{"command": "annotate", "ontology_iri": "https://example.org/module.owl"},
{"command": "convert", "output": "module.owl"}
]
}
Quality check:
{
"steps": [
{"command": "merge", "input": ["edit.owl"]},
{"command": "report", "fail_on": "ERROR", "output": "report.tsv"}
]
}
Global Options
All tools support these global options:
| Option | Description |
|---|---|
catalog |
XML catalog file for resolving imports |
prefixes |
JSON-LD prefix file |
add_prefix |
Add individual prefixes (list) |
noprefixes |
Disable default prefixes |
verbose |
Enable verbose logging |
strict |
Fail on unparsed triples |
xml_entities |
Use XML entities in RDF/XML output |
working_directory |
Working directory for file paths |
extra_args |
Escape hatch for rarely-used flags (list of strings) |
How to Contribute
- Fork the repository
- Create a feature branch from
main:git checkout -b feature/your-feature - Install dev dependencies:
uv sync --dev - Make your changes and ensure code quality:
uv run ruff format src/ uv run ruff check src/ - Commit your changes and push to your fork
- Open a Pull Request against
main
The main branch is protected — all changes must go through a PR.
License
BSD 2-Clause License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。