ai.nullary/nullary
ai.nullary/nullaryNegative results intelligence for drug discovery — query measured failures via MCP.
Tools · 35
Inactive small-molecule compound-target pairs.
Small molecules that failed selectivity.
Small-molecule ADMET failures.
Failed/ineffective CRISPR guides.
Non-dependency / failed essentiality screens.
Ancestry-specific CRISPR failures.
Antibody developability failures.
Discontinued/terminated clinical antibodies.
Failed peptide therapeutics.
Peptide stability/half-life failures.
PROTACs that failed degradation/ternary/permeability.
PROTAC E3-ligase recruitment / ternary failures.
ASOs/siRNAs that failed engagement/developability.
Oligonucleotide delivery failures.
Failed/terminated vaccines (by pathogen/indication).
Failed vaccine immunogen designs.
ADCs that failed at any stage.
ADC failures attributed to linker chemistry.
Bispecifics that failed at any stage.
Bispecific format/engineering failures.
ADMET failures across ALL modalities.
Drug-drug interaction failures.
Approaches that failed for a mechanism (by target).
Findings that failed to replicate.
Clinical/preclinical safety failures across modalities.
ALL failed approaches against a target across every modality.
ALL failed approaches for an indication across every modality.
Vaccine + antimicrobial + antibody failures for a pathogen.
A compound (structure, name, max clinical phase) + its full negative profile across modalities/sources.
Full provenance + detail for a single finding by id.
Target 'graveyard' / exhaustion index: how many distinct compounds/agents have been tried against a target and failed, broken down by modality and outcome. Answers 'how picked-over is this target?'. A…
Coverage browse: the most heavily-pursued ('graveyard') targets, ranked by recorded negative findings. Optional family filter (kinase, gpcr, protease, nuclear_receptor, ion_channel, transporter, phosp…
Summary of the Layer-1 inactivity-scoring model registry: how many per-target models, split by family, and median scaffold-split ROC-AUC.
Per-target Layer-1 model card: training counts and held-out scaffold-split metrics (ROC-AUC, PR-AUC, Brier, calibration). Accepts a gene symbol (e.g. EGFR) or UniProt accession (e.g. P00533).
Per-modality and per-source coverage stats (honest Phase-1 numbers).
Similar MCP servers embedding-nearest
How to use
Add to your Claude Desktop / Cursor / Cline MCP config:
{
"mcpServers": {
"ai.nullary/nullary": {
"url": "https://mcp.nullary.ai/mcp",
"transport": "streamable-http"
}
}
}