Typed Python SDK for the Modulector API. The package contains client helpers for querying predicted and experimentally validated miRNA target interactions, miRNA aliases, methylation sites, disease associations, drug associations, and PubMed subscriptions without installing the Django backend dependencies.
pip install modulector-sdkfrom modulector_sdk import (
get_mirna_details,
get_mirna_target_interactions,
get_mirna_target_validations,
)
details = get_mirna_details("hsa-miR-21-5p")
interactions = get_mirna_target_interactions(
mirna="hsa-miR-21-5p",
gene="PTEN",
include_pubmeds=True,
)
validations = get_mirna_target_validations(
mirna="hsa-miR-122-5p",
target="SLC7A1",
experiment="Western blot",
)Set MODULECTOR_API_BASE_URL to target a different Modulector deployment:
MODULECTOR_API_BASE_URL=https://your-modulector.example.org python script.pyEvery service function also accepts a base_url keyword argument for
per-request overrides.
The SDK also installs a Model Context Protocol server that exposes the same service functions as LLM-callable tools.
For local stdio MCP clients:
modulector-mcpExample client configuration:
{
"mcpServers": {
"modulector": {
"command": "modulector-mcp"
}
}
}For Streamable HTTP clients:
modulector-mcp --transport streamable-http --host 127.0.0.1 --port 8000Then connect the client to http://127.0.0.1:8000/mcp.
The server includes tools for predicted and experimentally validated miRNA target interactions, miRNA details and aliases, miRNA identifier resolution, methylation site lookup and annotation, disease associations, drug associations, and PubMed news subscriptions. Use MODULECTOR_API_BASE_URL or each tool's base_url argument to target a custom Modulector deployment.
When deployed with the root Docker Compose stack, nginx proxies the MCP
Streamable HTTP endpoint at https://<mydomain.com>/mcp.
Build this SDK from the sdk/ directory:
uv buildThe generated source distribution and wheel are written to sdk/dist/.