Predictive intelligence for drug discovery

Nullary is an MCP server providing structured, citable access to negative results in drug discovery.

CommunityOnlineNo sign-inGlobalFreeRead-only

What it can do

  • Search Inactive Compounds: Inactive small-molecule compound-target pairs.
  • Search Failed Selectivity: Small molecules that failed selectivity.
  • Search Admet Failures: Small-molecule ADMET failures.

What data it sees

Do you need an account

No: the server works without sign-in

Nullary is an MCP server providing structured, citable access to negative results in drug discovery. Every record carries full provenance (DOI, PMID, source URL, license). Covers 7 modalities: small molecules (84.5M records), CRISPR (37.6M), clinical trials (104K terminated), peptides, antibodies, PROTACs, and oligonucleotides.

Sources include ChEMBL, PubChem BioAssay, DepMap, BioGRID ORCS, GenomeCRISPR, ClinicalTrials.gov, EudraCT, Drugs@FDA, Thera-SAbDab, SAbDab, FLAb, PROTAC-DB, THPdb, AOBase, Retraction Watch, and 10 others. Fully cited.

How to connect

How to connect Predictive intelligence for drug discovery to Claude

The server address to paste into Claude settings:

https://drug-discovery.run.tools

  1. Open Claude (claude.ai or the desktop app).
  2. Go to Settings → Connectors.
  3. Click Add custom connector.
  4. Paste the server address copied below into Remote MCP server URL and click Add.
  5. In a chat, click + → Connectors and switch the new connector on.

Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

Authorization

After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

How to connect Predictive intelligence for drug discovery to ChatGPT

The server address to paste into ChatGPT settings:

https://drug-discovery.run.tools

  1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
  2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
  3. Open Settings → Connectors and click Create.
  4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
  5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
  6. In a new chat open + → Apps/Connectors and enable the connector.

OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

Authorization

On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

How to connect Predictive intelligence for drug discovery to Cursor

The server address to paste into Cursor settings:

https://drug-discovery.run.tools

Fastest: click Open in Cursor below and confirm the prompt.

Manually:

  1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
  2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
  3. Save the file. The server appears in the MCP list; authorize it there if asked.

Authorization

If the service needs sign-in, an authorize button appears next to the server in the MCP list.

Server tool list (35)

Raw names from tools/list. Only developers need these.

search_inactive_compoundsInactive small-molecule compound-target pairs.
search_failed_selectivitySmall molecules that failed selectivity.
search_admet_failuresSmall-molecule ADMET failures.
search_failed_guidesFailed/ineffective CRISPR guides.
search_failed_essentiality_screensNon-dependency / failed essentiality screens.
search_ancestry_specific_failuresAncestry-specific CRISPR failures.
search_developability_failuresAntibody developability failures.
search_failed_clinical_antibodiesDiscontinued/terminated clinical antibodies.
search_failed_peptide_therapeuticsFailed peptide therapeutics.
search_peptide_stability_issuesPeptide stability/half-life failures.
search_failed_protacsPROTACs that failed degradation/ternary/permeability.
search_protac_e3_issuesPROTAC E3-ligase recruitment / ternary failures.
search_failed_oligonucleotidesASOs/siRNAs that failed engagement/developability.
search_oligo_delivery_failuresOligonucleotide delivery failures.
search_failed_vaccinesFailed/terminated vaccines (by pathogen/indication).
search_vaccine_immunogenicity_failuresFailed vaccine immunogen designs.
search_failed_adcsADCs that failed at any stage.
search_adc_linker_failuresADC failures attributed to linker chemistry.
search_failed_bispecificsBispecifics that failed at any stage.
search_bispecific_format_failuresBispecific format/engineering failures.
search_admet_failures_all_modalitiesADMET failures across ALL modalities.
search_drug_drug_interaction_failuresDrug-drug interaction failures.
search_mechanism_failuresApproaches that failed for a mechanism (by target).
search_failed_replicationsFindings that failed to replicate.
search_safety_failuresClinical/preclinical safety failures across modalities.
search_target_historyALL failed approaches against a target across every modality.
search_indication_historyALL failed approaches for an indication across every modality.
search_pathogen_historyVaccine + antimicrobial + antibody failures for a pathogen.
get_compoundA compound (structure, name, max clinical phase) + its full negative profile across modalities/sources.
get_finding_provenanceFull provenance + detail for a single finding by id.
get_target_landscapeTarget '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?'. Accepts a gene symbol (e.g. EGFR) or a UniProt accession (e.g. P00533).
list_top_targetsCoverage browse: the most heavily-pursued ('graveyard') targets, ranked by recorded negative findings. Optional family filter (kinase, gpcr, protease, nuclear_receptor, ion_channel, transporter, phosphatase, other).
list_modelsSummary of the Layer-1 inactivity-scoring model registry: how many per-target models, split by family, and median scaffold-split ROC-AUC.
get_model_cardPer-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).
get_coveragePer-modality and per-source coverage stats (honest Phase-1 numbers).