Nullary MCP server

• Registry nameai.nullary/nullaryverified in the MCP registry

The registry name sits under nullary.ai. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

• Stateanswerschecked 2026-10-23
• Tools35
• Server calls itselfNullary
• Server version1.0.1
• Protocol version2025-06-18

On 2026-10-23 the server answered initialize and tools/list without a sign-in. First answer in 147 ms.

• Its tools do every use case we checked: Search failed compounds and terminated trials by target; Check whether a target's inhibitors previously failed; Review failed CRISPR screens before repeating experiments.
• Checked on 4 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 4 so far
• What an agent can do with it3 of 3 use cases
• Answeredmeasuring4 of the 7 daily checks it takes to print a share
• Median first answer151 ms
• Space in an agent's context≈ 7,304 tokensthe definitions of all its tools (35), which an agent carries on every turn
• Tool definitionsA · 100 of 100
• Typed results35 of 35 toolsdeclared outputSchema: results come typed, read once

• Every tool passes all six checks.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• search_inactive_compounds

  Inactive small-molecule compound-target pairs.

• search_failed_selectivity

  Small molecules that failed selectivity.

• search_admet_failures

  Small-molecule ADMET failures.

• search_failed_guides

  Failed/ineffective CRISPR guides.

• search_failed_essentiality_screens

  Non-dependency / failed essentiality screens.

• search_ancestry_specific_failures

  Ancestry-specific CRISPR failures.

• search_developability_failures

  Antibody developability failures.

• search_failed_clinical_antibodies

  Discontinued/terminated clinical antibodies.

• search_failed_peptide_therapeutics

  Failed peptide therapeutics.

• search_peptide_stability_issues

  Peptide stability/half-life failures.

• search_failed_protacs

  PROTACs that failed degradation/ternary/permeability.

• search_protac_e3_issues

  PROTAC E3-ligase recruitment / ternary failures.

• search_failed_oligonucleotides

  ASOs/siRNAs that failed engagement/developability.

• search_oligo_delivery_failures

  Oligonucleotide delivery failures.

• search_failed_vaccines

  Failed/terminated vaccines (by pathogen/indication).

• search_vaccine_immunogenicity_failures

  Failed vaccine immunogen designs.

• search_failed_adcs

  ADCs that failed at any stage.

• search_adc_linker_failures

  ADC failures attributed to linker chemistry.

• search_failed_bispecifics

  Bispecifics that failed at any stage.

• search_bispecific_format_failures

  Bispecific format/engineering failures.

• search_admet_failures_all_modalities

  ADMET failures across ALL modalities.

• search_drug_drug_interaction_failures

  Drug-drug interaction failures.

• search_mechanism_failures

  Approaches that failed for a mechanism (by target).

• search_failed_replications

  Findings that failed to replicate.

• search_safety_failures

  Clinical/preclinical safety failures across modalities.

• search_target_history

  ALL failed approaches against a target across every modality.

• search_indication_history

  ALL failed approaches for an indication across every modality.

• search_pathogen_history

  Vaccine + antimicrobial + antibody failures for a pathogen.

• get_compound

  A compound (structure, name, max clinical phase) + its full negative profile across modalities/sources.

• get_finding_provenance

  Full provenance + detail for a single finding by id.

• get_target_landscape

  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?'. Accepts a gene symbol (e.g. EGFR) or a…

• list_top_targets

  Coverage 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_models

  Summary of the Layer-1 inactivity-scoring model registry: how many per-target models, split by family, and median scaffold-split ROC-AUC.

• get_model_card

  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).

• get_coverage

  Per-modality and per-source coverage stats (honest Phase-1 numbers).