MCP Tools DirectoryAbout this integration
Pre-clinical drug-target intelligence: 60 curated targets, 1,969 counted partner genes, 12 tools
- Transport
- stdio
- Authentication
- unknown
- Initial setup
- credentials
- Runtime
- unattended
- Evidence
- documented
- Version
- 1.7.2
- Package
- mosaic-mcp
- Last compatibility test
- Not independently tested
Connect your agent
mosaic-mcpThe package contains the query layer but no knowledge-graph data. Operators must supply and govern their own PostgreSQL database; the hosted Mosaic service is a different connection and is not counted as this package. No database or endpoint was contacted.
Capabilities: Search and profile drug targets, Query compounds, literature, patents, and clinical-trial data, Manage owner-scoped watchlists and research workspaces
Connected profiles
Additional details
io.github.sourabhnk/mosaic-mcp
Source ↗ · Checked 2026-09-171.7.2
Source ↗ · Checked 2026-09-17CC0-1.0; package licenses are separate
Source ↗ · Checked 2026-09-17pypi
Source ↗ · Checked 2026-09-17DATABASE_URL
Source ↗ · Checked 2026-09-17mosaic-mcp
Source ↗ · Checked 2026-09-191.7.1
Source ↗ · Checked 2026-09-19The package launches a stdio-only MCP server through its Python console script and requires a caller-controlled PostgreSQL connection.
Source ↗ · Checked 2026-09-19DATABASE_URL is required for the caller's PostgreSQL instance and MOSAIC_API_KEY is optional for Pro tools; the combination of database credentials and an optional service key does not fit one native authentication class.
Source ↗ · Checked 2026-09-19Configure an MCP client to launch mosaic-mcp over stdio with the required DATABASE_URL for the operator's own PostgreSQL instance; the documented hosted SSE endpoint is separate from this artifact.
Source ↗ · Checked 2026-09-19Publisher documentation reviewed; package and integration endpoint not executed or independently security-audited.
Source ↗ · Checked 2026-09-19