Patsnap Chemical Molecular

Chemical intelligence for compound structure search, fragment analysis, and ADMET property prediction.

该来源不提供完整文件导出（国内平台多为平台内托管），仅存元数据与原链

模型生成摘要（rules/v1 · 2026-10-01 19:54）：Chemical intelligence for compound structure search, fragment analysis, and ADMET property prediction
这是模型对公开材料的总结，不是官方声明，请以原链内容为准。

接入信息

传输形态: http
鉴权方式: 需 API Key（探测：需要鉴权（HTTP 401））
端点: https://patsnap-chemical-molecular--openpatsnap.run.tools

该服务需要凭证，请按官方文档申请后替换占位符

{
  "mcpServers": {
    "Patsnap Chemical Molecular": {
      "headers": {
        "Authorization": "Bearer \u003cYOUR_KEY\u003e"
      },
      "url": "https://patsnap-chemical-molecular--openpatsnap.run.tools"
    }
  }
}

能力清单

工具	说明
ls_structure_search	Search chemical structures. Use this tool when the user wants to search compounds by chemical structure. The Python service only keeps the MCP request parameters and forwards the payload to the downstream chemical endpoint. Args: smiles: str, query structure in SMILES format, example "CC(=O)OC1=C(C=CC=C1)C(O)=O" type: str, search type, supports "EXT" and "SIM" include_multi_component: bool, whether to include multi-component compounds, valid for EXT type only, default true include_stereo: bool, whether to consider stereochemistry, default true include_charge: bool, whether to consider charge, default true include_isotope: bool, whether to consider isotopes, default true include_radical: bool, whether to consider radicals, default true threshold: float, similarity threshold required for "SIM", example 0.99 offset: int, pagination offset, example 0 limit: int, page size, example 20 Bad Case: Do not use `query` as a parameter for free-text search. This endpoint uses structured query payloads and does not accept a `query` field. Use `smiles` to provide a chemical structure for structure-based search. Returns: dict[str, Any]: Response body: { "total": 191, "items": [ { "inchi_key": "BSYNRYMUTXBXSQ-UHQGOLLYSA-N", "compound_name": "2-acetyloxy-3,4,5-trideuteriobenzoic acid", "formula": "C9H8O4", "similarity": 1.0 } ] }
ls_structure_fetch	Fetch chemical structure details in batch. Batch fetch full detail records by InChIKey. Args: inchi_keys: List[str]. InChIKey list used to fetch compound details. Maximum 100 identifiers per request. Returns: dict[str, Any]: Response body: { "items": [ { "inchi_key": "BSYNRYMUTXBXSQ-UHFFFAOYSA-N", "compound_name": "ASPIRIN", "molecular_formula": "C9H8O4", "canonical_smiles": "CC(=O)Oc1ccccc1C(=O)O" } ] }
ls_patent_structure_fetch	Fetch patent-related chemical structures. Batch fetch chemical structures associated with a patent by patent ID or patent number. Args: patent_id: str, patent ID used to fetch related structures. pn: str, patent number used to resolve patent ID before fetching structures. offset: int, pagination offset, starting from 0. limit: int, page size. Maximum 100. Returns: dict[str, Any]: Response body: { "total": 28, "items": [ { "inchi_key": "ANAIPYUSIMHBEL-UHFFFAOYSA-N", "compound_name": "dioxidanediide", "formula": "O2", "smiles": "[O-][O-]" } ] }
ls_admet_predict	Predict ADMET properties for a list of molecules. Use this tool when the user wants to evaluate drug-likeness or pharmacokinetic properties of small molecules. It calls the ADMET-AI model to predict absorption, distribution, metabolism, excretion, and toxicity properties for each input SMILES. Args: smiles: List[str]. List of SMILES molecular formula strings. Each element must be a valid SMILES string (e.g. "CC(=O)Oc1ccccc1C(=O)O"). Minimum 1, maximum 100 molecules per request. Returns: Dict[str, Any]: Prediction response. Schema: { "type": "object", "properties": { "predictions": { "type": "array", "description": "Per-molecule ADMET prediction results", "items": { "type": "object", "properties": { "smiles": {"type": "string", "description": "Input SMILES string"}, "properties": { "type": "object", "description": "Predicted ADMET property values (null if prediction failed for that property)", "properties": { "molecular_weight": {"type": "number", "nullable": true, "description": "Molecular weight (Da)"}, "logP": {"type": "number", "nullable": true, "description": "Lipophilicity (octanol-water partition coefficient)"}, "hydrogen_bond_acceptors": {"type": "number", "nullable": true, "description": "Number of hydrogen bond acceptors"}, "hydrogen_bond_donors": {"type": "number", "nullable": true, "description": "Number of hydrogen bond donors"}, "tpsa": {"type": "number", "nullable": true, "description": "Topological polar surface area (Å²)"}, "AMES": {"type": "number", "nullable": true, "description": "AMES mutagenicity probability (0–1)"}, "BBB_Martins": {"type": "number", "nullable": true, "description": "Blood-brain barrier permeability probability (0–1)"}, "Bioavailability_Ma": {"type": "number", "nullable": true, "description": "Oral bioavailability probability (0–1)"}, "CYP1A2_Veith": {"type": "number", "nullable": true, "description": "CYP1A2 inhibition probability (0–1)"}, "CYP2C19_Veith": {"type": "number", "nullable": true, "description": "CYP2C19 inhibition probability (0–1)"}, "CYP2C9_Substrate_CarbonMangels": {"type": "number", "nullable": true, "description": "CYP2C9 substrate probability (0–1)"}, "CYP2C9_Veith": {"type": "number", "nullable": true, "description": "CYP2C9 inhibition probability (0–1)"}, "CYP2D6_Substrate_CarbonMangels": {"type": "number", "nullable": true, "description": "CYP2D6 substrate probability (0–1)"}, "CYP2D6_Veith": {"type": "number", "nullable": true, "description": "CYP2D6 inhibition probability (0–1)"}, "CYP3A4_Substrate_CarbonMangels": {"type": "number", "nullable": true, "description": "CYP3A4 substrate probability (0–1)"}, "CYP3A4_Veith": {"type": "number", "nullable": true, "description": "CYP3A4 inhibition probability (0–1)"}, "hERG": {"type": "number", "nullable": true, "description": "hERG channel blockade probability (0–1, cardiac toxicity risk)"}, "Caco2_Wang": {"type": "number", "nullable": true, "description": "Caco-2 cell permeability (cm/s, log scale)"}, "Clearance_Hepatocyte_AZ": {"type": "number", "nullable": true, "description": "Hepatocyte clearance (mL/min/10^6 cells)"}, "Clearance_Microsome_AZ": {"type": "number", "nullable": true, "description": "Microsomal clearance (mL/min/g)"}, "Solubility_AqSolDB": {"type": "number", "nullable": true, "description": "Aqueous solubility (log mol/L)"}, "VDss_Lombardo": {"type": "number", "nullable": true, "description": "Volume of distribution at steady state (L/kg)"} } } } } } } } Good Case: ✔ ls_admet_predict(smiles=["CC(=O)Oc1ccccc1C(=O)O"]) <- single molecule ADMET evaluation ✔ ls_admet_predict(smiles=["CC(=O)Oc1ccccc1C(=O)O", "c1ccc(cc1)N"]) <- batch evaluation of multiple candidates Bad Case: ❌ ls_admet_predict(smiles=[]) <- empty list is not allowed ❌ ls_admet_predict(smiles=["not-a-smiles-string"]) <- invalid SMILES will cause prediction failure for that molecule
ls_chemical_mcs_analyze	Analyze fragment composition from multiple chemical structures. Use this tool when the user wants to analyze fragment distribution from a set of SMILES structures. It analyzes the frequent fragments and count how many times each fragment appears across the input structures. Args: structures: List[str]. List of SMILES structures to analyze. Must contain at least one valid SMILES string (e.g. ["CCO", "CC(C)O", "c1ccccc1"]). Returns: Dict[str, Any]: MCS analysis response. Schema: { "type": "object", "properties": { "fragments": { "type": "array", "description": "Fragment list with occurrence counts", "items": { "type": "object", "properties": { "fragment": {"type": "string", "description": "SMILES of the fragment structure"}, "count": {"type": "integer", "description": "Number of times this fragment appears across input structures"} } } } } } Good Case: ✔ ls_chemical_mcs_analyze(structures=["CCO", "CC(C)O", "CCCO"]) <- find fragments among a series of alcohol structures Bad Case: ❌ ls_chemical_mcs_analyze(structures=[]) <- empty list is not allowed ❌ ls_chemical_mcs_analyze(structures=["not-a-smiles"]) <- invalid SMILES will cause analysis failure

Patsnap Chemical Molecular

接入信息

能力清单

相关资产

ia-qa.com/mcp llm and RAG testing - Dev/QA toolbox

AlpineLead-OI

Grimoire Spellbook Server

Shipping Rates

Freightgate

Document to Markdown & RAG Chunks