CovaSyn Chemistry Tools
Server Details
Chemistry, pharma and biotech tools for AI assistants: NMR, MS, IR and UV analysis, ADMET and ICH M7, stability (ICH Q1), HPLC methods, DoE, docking, retrosynthesis, biologics. Deterministic results you can check. Built in Leipzig, Germany.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- Repository
- oliverkraft93-ops/covasyn-mcp-examples
- GitHub Stars
- 0
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 134 tools
Tools are grouped by domain prefixes and many descriptions explicitly say which tool to use, but there are real overlaps: covabasic_druglikeness vs covachem_druglikeness, covabasic_analyze vs covachem_profile, and multiple NMR/MS identification, verification, and ranking tools. An agent can often infer the right choice, but misselection remains plausible.
All tool names use lower snake_case with a consistent domain prefix such as covabasic_, covabio_, covadoe_, and covnmr_. Most append an action, though some are noun phrases like covabasic_atom_indices, covabasic_functional_groups, and covadoe_design_space, so the verb_noun pattern is not universal.
At 134 tools, the surface is far beyond a well-scoped set and creates a severe selection and maintenance burden. Even as a broad chemistry platform, the number of overlapping specialized tools and absorbed/retired aliases makes the count extreme rather than merely heavy.
Coverage is very broad across cheminformatics, biosequence analysis, DoE, IR, MS, NMR, UV-Vis, solubility, and toxicity. However, there are obvious dead references to tools not present in the set, such as covachem_pka, covachem_adme, and covafold examples in the job tools, which will cause agent dead ends.
Available Tools
134 toolscovabasic_analyzeCInspect
Compute molecular descriptors from SMILES.
scope options:
"basic" — 40+ RDKit descriptors (atoms, bonds, rings, scaffolds, HBD/HBA, MW, complexity)
"extended" — basic + Chi indices, Kier-Hall, Gasteiger charges, flexibility ratio
"lipophilicity" — LogP, MR, TPSA, LabuteASA surface descriptors
"mordred" — 100 Mordred 2D descriptors (autocorrelation, topology, information content)
"all" — everything combined
Inputs: smiles (required), scope, api_key. Returns: result
covabasic_atom_indicesAInspect
Return atom indices for a named structural query — replaces hand-written SMARTS.
Use this when you need atom positions for a structural feature
(e.g. "which atoms are in the aromatic ring?", "where are the stereocenters?",
"which carbon has the highest oxidation state?").
Returns: `{indices: [...], groups: [[...]], smarts_used: "..."}`. Plus:
- `named_groups: [{name, atoms, smarts}]` for `functional_group_atoms`
- `oxidation_states: [{atom, state}]` for `oxidation_state`
Supported query names (22):
ring/topology : aromatic_ring, aliphatic_ring, saturated_ring,
heterocycle, fused_ring, bridgehead
atom-class : halogen_atom, carbon_atom, heavy_atom, csp3_carbon
stereo : stereocenter, r_s_stereocenter, unspecified_stereocenter
bonds & chains : rotatable_bond, branch_point, chain_termini,
longest_carbon_chain
derived : murcko_scaffold_atoms, brics_break_atoms,
functional_group_atoms
H-bonding / oxidation:
hba (Lipinski-style hydrogen-bond acceptors),
oxidation_state (per-carbon Pauling-EN ox-state,
with parallel `oxidation_states` field)
Examples:
query="aromatic_ring" + smiles="c1ccccc1" → indices [0..5], one group
query="r_s_stereocenter" + smiles="C[C@H](O)C(=O)O" → stereo carbon indices
query="hba" + smiles="CC(=O)O" → carbonyl + hydroxyl O indices
query="oxidation_state" + smiles="CC(=O)O" → carbon indices PLUS
oxidation_states=[{atom:0, state:-3}, {atom:1, state:+3}]
query="functional_group_atoms" + any SMILES → also returns named_groups list
Cost: 2 credits, available on Free tier.Inputs: query (required), smiles (required), api_key. Returns: result
covabasic_bricsBInspect
BRICS retrosynthetic fragmentation into purchasable building blocks. Returns fragment SMILES and bond-break locations.
Inputs: smiles (required), api_key. Returns: result
covabasic_canonicalizeBInspect
Return the RDKit canonical SMILES. Useful to normalize SMILES before comparison or database lookup.
Inputs: smiles (required), api_key. Returns: result
covabasic_convertAInspect
Unit conversion for chemistry (30+ conversions). Examples: "mol_to_g", "g_to_mol", "ug_to_mg", "celsius_to_kelvin", "ppm_to_molarity", "atm_to_pa". Pass mw for mass-mole conversions.
Inputs: value (required), conversion (required), mw, api_key. Returns: result
covabasic_convert_formatBInspect
Convert between chemical structure formats. Auto-detects input format.
Supported input: SMILES, InChI, MOL block, SDF block
to_format options: "smiles", "inchi", "inchikey", "molblock", "sdf"
Inputs: input_string (required), api_key, to_format. Returns: result
covabasic_depictBInspect
Generate 2D molecule depiction(s). Returns SVG string or base64-encoded PNG.
Single molecule: provide smiles. Add highlight_smarts to highlight a substructure.
Grid of molecules: provide smiles_list (list of SMILES) + optional legends.
img_format: "svg" (default, scalable) or "png" (base64-encoded).
Inputs: width, height, smiles, api_key, +5 more. Returns: result
covabasic_describe_structureAInspect
Explain what a structure IS, in prose — parent framework, ring system, functional groups, substituents.
Turns a SMILES into a readable chemical description rather than a table of
numbers. Useful for reports, for captioning a drawn structure, and for
checking that a structure is what you meant to draw.
style="human" — compact, chemistry-facing narrative: the parent skeleton,
heteroatom positions, unsaturation, the principal
characteristic group, then substituents with their locants.
style="technical" — the full naming walkthrough, including how the name was
assembled token by token.
Returns: `{name, paragraphs: [...], text, rules_hit, smiles, style, engine}`,
or `{error, smiles}`. `text` is the paragraphs joined for direct display.Inputs: smiles (required), style, verify, api_key. Returns: result
covabasic_druglikenessBInspect
Evaluate drug-likeness: Lipinski Ro5, Veber, Ghose, Egan, Muegge, Ro3 (fragments), QED score, synthetic accessibility (SA) score.
Inputs: smiles (required), api_key. Returns: result
covabasic_equationAInspect
Solve a physicochemical equation. Pass equation name + parameter dict.
Equations and their params:
"arrhenius" — {k1, t1_k, k2?, t2_k?, ea_kj?} compute Ea or predict k at new T
"vant_hoff" — {k1, t1_k, k2, t2_k} compute reaction enthalpy delta-H
"henderson_hasselbalch" — {pka, ph, acid (bool)} compute ionization fractions
"ideal_gas" — {n?, p_atm?, v_l?, t_k?} PV=nRT, solve for missing variable
"beer_lambert" — {absorbance?, epsilon?, concentration?, path_length?} A=ecl
"dilution" — {c1, v1, c2?, v2?} C1V1=C2V2, solve for missing variable
Inputs: params (required), equation (required), api_key. Returns: result
covabasic_filter_checkAInspect
Screen for problematic substructures: PAINS (pan-assay interference), BRENK (unwanted groups), NIH MLPCN filters, ZINC druglike filters. Returns alert count + details per filter set.
Inputs: smiles (required), api_key. Returns: result
covabasic_fingerprintCInspect
Generate molecular fingerprints: Morgan (ECFP4), MACCS keys, RDKit topological. Returns bit vectors as hex strings.
Inputs: smiles (required), api_key. Returns: result
covabasic_functional_groupsCInspect
Count ~65 functional groups: alcohols, phenols, amines, amides, esters, ethers, ketones, aldehydes, carboxylic acids, sulfonamides, nitro, halogens, heterocycles, etc.
Inputs: smiles (required), api_key. Returns: result
covabasic_green_chemistryAInspect
Calculate green chemistry metrics: E-factor (kg waste / kg product) and/or PMI (Process Mass Intensity = total input / product).
Inputs: mass_product_kg (required), api_key, mass_waste_kg, mass_all_inputs_kg. Returns: result
covabasic_image_to_smilesAInspect
Convert a chemical structure image to SMILES via Mathpix OCR. Input: base64-encoded PNG/JPG of a drawn molecule. Returns SMILES, confidence, and all detected structures. Requires Mathpix API credentials on server.
Inputs: image_base64 (required), api_key. Returns: result
covabasic_iupac_nameAInspect
Get the IUPAC name of a structure — generating one when it is unknown to any database.
Two sources, tried in order, and the response always says which one answered:
- `source="pubchem"` — the curated, human-checked IUPAC name. Authoritative,
but only exists for compounds PubChem indexes.
- `source="openclatura"` — a name constructed from the molecular graph by a
deterministic IUPAC Blue Book (2013) rule engine. Works on any valid
structure, including novel compounds, intermediates and analogues.
Parameters:
prefer_pubchem — check PubChem first (default true). Set false to force
generation, or to keep the call purely offline and ~100x faster: the
rule engine takes ~5 ms, the PubChem round-trip 0.6-1 s.
explain — also return `rules_hit`, the Blue Book rules that fired
(e.g. P-44, P-61). Costs ~5x the naming time; off by default.
verify — round-trip the generated name through OPSIN and report whether it
parses back to the same structure. Needs a Java runtime; without one the
result reports `verification="skipped_no_java"` instead of failing.
Returns: `{name, smiles, source, generated, engine}` (+ `rules_hit`,
`verification` when requested), or `{error, smiles}`.
A generated name is a systematic construction, not a database record —
treat `generated=true` accordingly.Inputs: smiles (required), verify, api_key, explain, +1 more. Returns: result
covabasic_name_to_structureAInspect
Turn systematic chemical names into structures (OPSIN) — the inverse of covabasic_iupac_name.
Parses up to 50 English IUPAC / systematic names per call into canonical
SMILES and InChIKey with OPSIN, a deterministic grammar-based parser. It
does not look anything up: trade names, abbreviations, typos and
non-English names come back with `smiles=null` rather than a guess.
Translate a German or Chinese name to English first.
Returns: `{results: [{name, smiles, inchikey}], engine}`, or
`{error, reason}` (reason "no_java" when the server has no Java runtime).Inputs: names (required), api_key. Returns: result
covabasic_pubchemBInspect
Query PubChem database. NETWORK REQUIRED.
action options:
"lookup" — find compound by name, CAS, SMILES, or InChIKey. Returns CID, IUPAC name, InChI, MW, XLogP, TPSA.
"smiles" — resolve a compound name or CAS to canonical SMILES.
"properties" — fetch computed properties for a PubChem CID (pass CID as query).
"similarity" — find similar compounds by Tanimoto (pass SMILES as query, set threshold 0-1).
Inputs: query (required), action, api_key, threshold, +2 more. Returns: result
covabasic_reynoldsBInspect
Reynolds number calculation with flow regime classification (laminar / transitional / turbulent). For pipe flow and mixing design.
Inputs: diameter_m (required), velocity_m_s (required), density_kg_m3 (required), viscosity_pa_s (required), +1 more. Returns: result
covabasic_scaffold_analysisBInspect
Murcko scaffold decomposition: core ring framework, side chains, generic scaffold, ring systems, linkers. Useful for series analysis and SAR.
Inputs: smiles (required), api_key. Returns: result
covabasic_similarityBInspect
Compare two molecules by structural similarity.
method="all" — Tanimoto, Dice, Cosine, MCS (Maximum Common Substructure) method="tversky" — asymmetric Tversky index (alpha/beta control substructure vs superstructure bias)
Inputs: smiles_a (required), smiles_b (required), beta, alpha, +2 more. Returns: result
covabasic_standardizeAInspect
Standardize a SMILES string.
mode="basic" — salt stripping, neutralization, canonical tautomer mode="enhanced" — + charge-parent extraction, metal disconnection, stereo normalization
Inputs: smiles (required), mode, api_key. Returns: result
covabasic_stoichiometryCInspect
Calculate reactant masses needed for a target product mass. Input: MW of product, list of reactant MWs, stoichiometric equivalents, and target mass in grams.
Inputs: mw_product (required), equivalents (required), mw_reactants (required), target_mass_g (required), +1 more. Returns: result
covabasic_substructureCInspect
Search for a SMARTS substructure pattern in a molecule. Returns match count, atom indices of each hit, and highlighted atoms.
Inputs: smiles (required), pattern_smarts (required), api_key. Returns: result
covabasic_tautomersBInspect
Enumerate tautomers of a molecule (keto-enol, amide-imidol, etc.). Returns canonical tautomer + all alternatives.
Inputs: smiles (required), api_key, max_tautomers. Returns: result
covabasic_validateBInspect
Validate SMILES with 15+ checks: parsability, valence, stereo, aromaticity, charge, size limits, isotopes, radicals.
Inputs: smiles (required), api_key. Returns: result
covabio_adcBInspect
Design an antibody-drug conjugate (ADC): estimate DAR (drug-to-antibody ratio), total molecular weight, conjugatable cysteine/lysine sites, aggregation risk, and payload hydrophobicity impact.
Parameters:
antibody_sequence: Antibody amino acid sequence.
payload_smiles: Drug payload SMILES (e.g. MMAE, DM1).
target_dar: Target drug-to-antibody ratio (default 4.0).
Returns: {total_mw, dar, conjugation_sites, aggregation_risk, payload_properties}Inputs: payload_smiles (required), antibody_sequence (required), api_key, target_dar. Returns: result
covabio_alignBInspect
Pairwise sequence alignment using BioPython. Supports BLOSUM62, BLOSUM45, BLOSUM80 substitution matrices. Returns alignment score, identity percentage, aligned sequences, and gap statistics.
Parameters:
seq1: First amino acid sequence.
seq2: Second amino acid sequence.
matrix: Scoring matrix — "BLOSUM62" (default), "BLOSUM45", "BLOSUM80".
Returns: {score, identity_pct, aligned_seq1, aligned_seq2, n_gaps}Inputs: seq1 (required), seq2 (required), matrix, api_key. Returns: result
covabio_antibodyAInspect
Identify CDR regions (complementarity-determining regions) and framework regions in an antibody sequence. Supports Kabat, Chothia, and IMGT numbering schemes. Returns CDR boundaries, chain type (heavy/light), framework sequences, and CDR developability flags.
Parameters:
sequence: Antibody heavy or light chain amino acid sequence.
scheme: Numbering scheme — "kabat" (default), "chothia", or "imgt".
Returns: {chain_type, CDR1/2/3 sequences and positions, framework_regions, cdr_developability}Inputs: sequence (required), scheme, api_key. Returns: result
covabio_batch_profileAInspect
Profile up to 100 protein sequences in one call — the same 30+ biophysical properties as covabio_protein_profile, per sequence. Charged once per call, not per sequence. Use for candidate panels, variant series, or a whole FASTA out of covabio_parse_fasta.
A sequence that fails does not abort the batch: its slot carries {error: true, message, sequence} and the rest still return. Over 100 sequences the call is rejected rather than truncated, so a silently short result is not possible — split the batch.
Parameters:
sequences: List of amino acid sequences (max 100).
Returns: {results: [profile | {error, message, sequence}]} in input orderInputs: sequences (required), api_key. Returns: result
covabio_charge_profileAInspect
Compute net charge versus pH over a chosen window, plus the isoelectric point. Use for buffer and formulation-pH selection, for predicting IEX binding and elution behaviour, and for reading a pI shift between a reference and a comparability lot.
Parameters:
sequence: Protein amino acid sequence.
ph_start: Lower bound of the pH window (default 2.0).
ph_end: Upper bound of the pH window (default 12.0).
Returns: {charge_profile: [{ph, charge}], isoelectric_point}Inputs: sequence (required), ph_end, api_key, ph_start. Returns: result
covabio_cleavage_sitesAInspect
Map protease cleavage sites and score proteolytic stability. Covers trypsin (after K/R, not before P), chymotrypsin (after F/W/Y, not before P), pepsin (after F/L) and elastase (after A/G/S/V). Returns per-protease position lists, a total site count, a 0-100 stability score and a classification (highly_stable / moderate / proteolytically_labile).
Two uses: predicting in-vivo half-life and serum stability for a peptide therapeutic, and planning the digest for a peptide-mapping or MAM method — the trypsin list is the peptide map you will get.
Returns the same protease analysis that covabio_peptide embeds, on its own and for any protein length.
Parameters:
sequence: Amino acid sequence.
Returns: {sites: {trypsin, chymotrypsin, pepsin, elastase}, total_cleavage_sites, stability_score, classification}Inputs: sequence (required), api_key. Returns: result
covabio_codon_adaptationAInspect
Compute the Codon Adaptation Index (CAI) of a DNA coding sequence against host codon usage. Returns the score, how many codons were scored, how many hit the host's optimal codon, and a classification (highly_optimized >0.8 / well_optimized >0.6 / moderately_optimized >0.4 / poorly_optimized). Use to judge an expression construct before ordering synthesis, or to compare two vendor optimizations.
Scored against the optimal codon per amino acid, so it measures optimal-codon fraction rather than the geometric-mean CAI of Sharp & Li 1987. Stop codons are excluded. For mRNA therapeutics use covabio_mrna, which adds GC, CpG and uridine content.
Parameters:
dna_sequence: Coding DNA sequence (ATG...stop).
organism: Host codon table — "ecoli" (default) or "human".
Returns: {cai_score, total_codons, optimal_codons, scored_codons, organism, classification}Inputs: dna_sequence (required), api_key, organism. Returns: result
covabio_deamidationAInspect
Locate asparagine deamidation hotspots in a protein sequence — the single most cited chemical-degradation CQA for antibodies and fusion proteins. Scans for the six Asn-Xaa motifs ordered by known lability: NG (critical), NS (high), NH and ND (medium), NA and NT (low). Use to justify a deamidation-monitoring strategy, to compare candidate sequences before lead selection, or to explain a charge-variant shift observed on IEX/icIEF.
This is a sequence-motif rule set, not a structural or kinetic prediction: solvent exposure, local flexibility, pH and temperature all modulate the real rate, and a buried NG can be stable while an exposed ND is not. Treat the output as a site list to test, not a rate.
Parameters:
sequence: Protein amino acid sequence (1-letter code).
Returns: {deamidation_sites: [{position, motif, risk, context}]} sorted by positionInputs: sequence (required), api_key. Returns: result
covabio_developabilityAInspect
Assess biopharmaceutical developability across 8 risk dimensions: aggregation propensity, deamidation hotspots, oxidation sites, glycosylation risk, charge heterogeneity, hydrophobic patches, sequence liabilities, and post-translational modification risk. Returns composite score (0-100) with per-dimension breakdown and flagged positions.
Parameters:
sequence: Protein/antibody amino acid sequence.
metal_binding_context: Set True for chelators/metal-binding constructs so functional
Cys/His/carboxylate residues aren't penalized as liabilities (default False).
Returns: {composite_score, risk_dimensions: {aggregation, deamidation, ...}, flagged_positions}Inputs: sequence (required), api_key, metal_binding_context. Returns: result
covabio_immunogenicityAInspect
Screen protein sequence for potential T-cell epitopes (MHC-II binding). Scans 9-mer windows against 8 common HLA-DRB1 alleles using positional scoring matrices. A window counts as an epitope when it clears an absolute, background-calibrated score threshold (each allele flags at most 5% of random 9-mers of natural composition), so epitope_density is comparable between sequences: a sequence of random natural composition scores ~0.09 against this panel. Deimmunization suggestions target the P1/P4/P6/P9 anchor positions that drive binding. This is an in-silico binding prediction from hand-curated matrices, NOT an ADA rate and not calibrated against measured affinities.
Parameters:
sequence: Protein amino acid sequence.
alleles: Optional HLA allele list (default: 8 common DRB1 alleles).
Returns: {n_epitopes, epitopes: [{position, peptide, allele, score, binding_alleles}], epitope_density, risk_classification (low <0.18, moderate <0.27, high), hotspots, per_allele_scores, deimmunization_suggestions: [{position, original_aa, suggested_aa, position_class, predicted_reduction}], binder_thresholds, threshold_basis}Inputs: sequence (required), alleles, api_key. Returns: result
covabio_mam_reportAInspect
Consolidated Multi-Attribute Method hotspot report: every sequence-liable modification site in one ranked list, across seven modification types — deamidation, oxidation, isomerization, N-glycosylation sequons, acid-labile cleavage, N-terminal pyroglutamate, and C-terminal lysine clipping. This is the site list a MAM peptide-mapping method has to cover, and the starting point for an attribute-criticality assessment.
Sequence-derived only: it tells you which residues CAN carry an attribute, not which ones DO in a given lot. It replaces neither the LC-MS measurement nor the criticality ranking.
Parameters:
sequence: Protein/antibody amino acid sequence.
Returns: {hotspots: [{position, modification_type, motif, risk, context}]}Inputs: sequence (required), api_key. Returns: result
covabio_mrnaBInspect
Evaluate mRNA therapeutic sequence: codon adaptation index (CAI), GC content optimization, CpG dinucleotide analysis, uridine content, and overall optimization score for translational efficiency.
Parameters:
coding_sequence: mRNA coding DNA sequence (ATG...stop).
organism: Target organism — "human" (default) or "ecoli".
Returns: {cai_score, gc_content, cpg_count, uridine_fraction, optimization_score}Inputs: coding_sequence (required), api_key, organism. Returns: result
covabio_oligoAInspect
Compute oligonucleotide properties: melting temperature (Tm via SantaLucia 1998 nearest-neighbor), GC content, molecular weight, extinction coefficient, hairpin detection, and self-complementarity check.
Parameters:
sequence: Nucleotide sequence (e.g. "ATCGATCG").
oligo_type: "DNA" (default) or "RNA".
Returns: {tm_celsius, gc_content, molecular_weight, hairpin_detected, self_complementary}Inputs: sequence (required), api_key, oligo_type. Returns: result
covabio_parse_fastaAInspect
Parse FASTA text into named sequence entries. Handles multi-line sequences, multiple records, and raw sequence without a header (auto-named seq_1, seq_2, ...). Use to turn a pasted or downloaded FASTA into the input for covabio_batch_profile or any single-sequence tool.
Parameters:
fasta_text: FASTA-formatted text (one or more records).
Returns: {entries: [{name, sequence}]}Inputs: fasta_text (required), api_key. Returns: result
covabio_peptideAInspect
Profile a peptide sequence: biophysical properties plus peptide-specific analyses including cell-penetrating peptide (CPP) score, antimicrobial peptide (AMP) prediction, proteolytic stability (trypsin/chymotrypsin/pepsin cleavage sites), and N-end rule half-life.
Parameters:
sequence: Peptide amino acid sequence (typically 5-50 residues).
Returns: {properties, cpp_score, amp_score, protease_sites, half_life}Inputs: sequence (required), api_key. Returns: result
covabio_protein_profileBInspect
Compute 30+ biophysical properties of a protein from its amino acid sequence. Returns molecular weight, isoelectric point, extinction coefficient, GRAVY (hydrophobicity), instability index, aromaticity, charge profile across pH range, amino acid composition, and secondary structure fractions.
Parameters:
sequence: Amino acid sequence (1-letter code, e.g. "MKWVTFISLLLLFSSAYS...").
Returns: {molecular_weight, isoelectric_point, gravy, instability_index, charge_profile, ...}Inputs: sequence (required), api_key. Returns: result
covabio_sequence_utilsAInspect
Sequence toolbox for construct design: 1-letter to 3-letter code conversion, codon table, reverse translation optimized for E. coli and for human codon usage, signal-peptide check, simple domain search, and restriction-enzyme sites found in the reverse-translated DNA. Use when going from a protein sequence to an expression construct.
Parameters:
sequence: Protein amino acid sequence.
Returns: {length, one_to_three, codon_table, reverse_translate_ecoli, reverse_translate_human, signal_peptide_check, domain_search, restriction_sites_in_dna}Inputs: sequence (required), api_key. Returns: result
covabio_sirnaAInspect
Design siRNA candidates against a target mRNA sequence using Reynolds 2004 rules. Scores 19-mer windows for efficacy, GC content, thermodynamic asymmetry, and off-target risk. Returns ranked candidate list.
Parameters:
target_sequence: Target mRNA sequence (DNA, minimum 21 nt).
Returns: {n_candidates, candidates: [{position, sense, antisense, score, gc_content}]}Inputs: target_sequence (required), api_key. Returns: result
covabio_viscosityAInspect
Estimate viscosity of a protein solution at high concentration. Uses sequence-based heuristic model (Sharma 2014) considering charge, hydrophobicity, and molecular weight. Critical for antibody formulation development.
Parameters:
sequence: Protein amino acid sequence.
concentration_mg_ml: Protein concentration (default 150 mg/mL for subcutaneous).
ph: Formulation pH (default 6.0).
Returns: {viscosity_cP, risk_level, contributing_factors}Inputs: sequence (required), ph, api_key, concentration_mg_ml. Returns: result
covachem_compareCInspect
Compare 2-5 molecules side by side.
Inputs: smiles_list (required), api_key. Returns: result
covachem_druglikenessCInspect
Evaluate drug-likeness filters (Lipinski, Ghose, Veber, PAINS, etc.).
Inputs: smiles (required), api_key. Returns: result
covachem_profileAInspect
Broad physicochemical scan — 80+ properties in one call (logP, TPSA, H-bond donors/acceptors, rotatable bonds, druglikeness flags, heuristic logS, etc).
Use this when the user wants a general overview of a molecule.
Do NOT use this if the user asks specifically about:
• SOLUBILITY in a solvent → use `covasolve_predict` (ML model with
confidence intervals, ~10× more accurate than this tool's heuristic
logS, and cheaper at 5 credits vs 6).
• DRUGLIKENESS only → `covachem_druglikeness` (1 cr).
• pKa values → `covachem_pka` (5 cr).
• ADME parameters only → `covachem_adme` (15 cr).
Cost: 6 credits (broad-scan premium). The targeted tools above are 1-5 cr
each and produce better results within their domain.Inputs: smiles (required), api_key. Returns: result
covadoe_alias_structureCInspect
Compute alias/confounding structure and design resolution.
Inputs: data (required), factor_columns (required), api_key, max_order. Returns: result
covadoe_analyzeAInspect
Run ANOVA analysis on experimental results.
Returns OLS coefficients (coded + uncoded, with a plain-language model_equation_uncoded),
Type III ANOVA (+ Type I sequential SS under anova_type1), and significance. Blocking and
split-plot terms are picked up automatically from "Block"/"WholePlot" columns. Optional:
`factor_info` (list of {name, factor_type:"categorical", levels:[...]}) enables multi-df categorical
F-tests; `response_family` ("gaussian"|"binomial"|"poisson") fits a GLM for non-normal responses.Inputs: data (required), response (required), api_key, factor_info, +3 more. Returns: result
covadoe_augment_designCInspect
Augment a screening design with RSM runs (axial, fold-over, BBD).
Inputs: existing_design (required), selected_factors (required), api_key, target_type, +2 more. Returns: result
covadoe_canonical_analysisCInspect
Canonical analysis: eigenvalues, stationary point, surface nature.
Inputs: model (required), api_key. Returns: result
covadoe_compare_designsCInspect
Compare and rank candidate designs for given factors.
Inputs: factors (required), goal, api_key, max_runs. Returns: result
covadoe_create_designAInspect
Create an experimental design (factorial, fractional, CCD, BBD, D/I-optimal, Taguchi, LHS, simplex mixture, etc.).
Optional: `constraints` — list of {terms:{factor:coef}, op:"<="|">="|"==", rhs} in REAL units —
filters the D/I-optimal candidate set. `mixture=True` with a simplex `design_type`
("simplex_lattice"/"simplex_centroid") builds a bounded-mixture design (L-pseudocomponent,
components sum to 1, each ≥ its `low`); `degree` sets the simplex-lattice degree; `n_levels` sets
the D/I-optimal candidate-grid density.
Model-aware optimal designs: `model_terms` ("linear"|"interaction"|"quadratic") sets the model
matrix the D/I-optimal criterion optimizes for. `n_blocks` (>1) adds a "Block" column, confounding
the highest-order interaction(s) for factorial designs. `n_runs` / `resolution` / `generators`
control the `fractional_factorial` fraction, in that precedence order: `n_runs` (a power of 2,
counting base runs before center points) pins the fraction exactly — e.g. n_runs=16 on 5 factors
gives 2^(5-1), resolution V; `resolution` (3/4/5...) picks the smallest fraction reaching it;
`generators` gives the defining generators verbatim as pyDOE2 tokens, one per factor
(e.g. ["a","b","c","ab","ac"] for 2^(5-2)).
`design_type="taguchi"` builds an orthogonal-array design (`taguchi_levels` 2 or 3). `hard_to_change`
(list of factor names) produces a split-plot layout with a "WholePlot" column. `process_factors`
(with `mixture=True`) builds a mixture-process crossed design (`process_design_type`,
`process_center_points`). `n_runs` also sets the run count for D/I-optimal designs (default
len(factors)*3+1) and the sample count for latin_hypercube / sobol_sequence.
CCD variant: `face` — "circumscribed" (default), "faced" (star points on the faces, α = 1, for
factors that cannot physically exceed their range) or "inscribed" (the CCC scaled so the star
points land on the stated range). `axial_alpha` sets the axial distance; the default is
ROTATABLE, (2**k)**0.25, and "orthogonal" gives the α where Σx² = n_runs. The choice is worth
several points of power — check it with `covadoe_design_power` before running the design.
Common `design_type` aliases resolve ("full_factorial", "ccd", "dsd", "bbd"); the response
reports the canonical name actually built.Inputs: factors (required), design_type (required), face, degree, +17 more. Returns: result
covadoe_design_powerAInspect
Planning-time statistical power from a design's REAL model matrix — no data needed.
Builds the design and takes `ncp = effect_size² / (X'X)⁻¹_jj`, so it is correct for centre
points, axial points and unbalanced designs — unlike `covadoe_power_analysis`, which assumes
orthogonality and overstates power on a CCD, Box-Behnken or DSD. Use this when planning an
experiment; use `covadoe_post_hoc_power` only once you have measured responses.
`design_type` + `n_factors` builds the design (same generator as `covadoe_create_design`, so the
answer describes the design you would actually get). Common aliases resolve ("full_factorial",
"ccd", "dsd"). Alternatively pass `design`, an explicit coded run matrix, to score a design you
already hold — D/I-optimal, mixture, split-plot or hand-built. `n_runs` / `model_terms` are
forwarded to the generator for the optimal designs.
`model`: "linear" | "interaction" | "quadratic" | "quadratic_no_interaction" (a DSD supports only
the last). `term` names the coefficient — "x1", "x1^2", "x1*x2"; omit it to get power for every
term and group, which differ sharply on the same design. `alpha` is the significance level;
the CCD axial distance is `axial_alpha`, and the two are easy to confuse.
CCD variant: `axial_alpha` defaults to ROTATABLE, (2**n_factors)**0.25 — the design JMP,
Design-Expert and Minitab hand you. Pass a number, or "orthogonal" for the α where Σx² = n_runs
(at 3 factors: main-effect power 0.9400 against the rotatable 0.8846 on the same 17 runs; the
orthogonal α is also the premise of `covadoe_power_analysis`, so on that design the two agree
for a main effect and disagree for every other term). `face` — "circumscribed" (default),
"faced" (star points on the faces, α = 1) or "inscribed"; a bool works, and an unknown value is
rejected rather than quietly treated as circumscribed.
Returns power, power_by_term, power_by_group, df_error, relative_variance ((X'X)⁻¹_jj), the
resolved design_summary (the α and factorial level actually built — read both, an inscribed CCD
always reports α = 1), and `orthogonal_shortcut_power` — what `covadoe_power_analysis` would
have claimed — with its signed error (signed: the shortcut does not always overstate). Returns
`estimable: false` plus a reason when the model cannot be fitted on the design.Inputs: face, term, alpha, model, +10 more. Returns: result
covadoe_design_spaceCInspect
Compute ICH Q8 design space (NOR, PAR, probability of meeting specs).
Inputs: data (required), response_specs (required), api_key, n_samples, +2 more. Returns: result
covadoe_find_optimumCInspect
Find optimal factor settings for a single response.
Inputs: model (required), goal, api_key. Returns: result
covadoe_fit_rsmCInspect
Fit a response surface model (quadratic RSM).
`factor_info` (list of {name, factor_type:"categorical", levels:[...]}) dummy-codes
categorical factors; without it every factor is fitted as numeric.Inputs: data (required), response (required), api_key, factor_info, +2 more. Returns: result
covadoe_model_selectionCInspect
Select best model terms via stepwise, all-subsets, or hierarchical selection.
Inputs: data (required), response (required), method, api_key, +1 more. Returns: result
covadoe_optimizeCInspect
Multi-response optimization (desirability, Pareto, weighted sum).
`factor_info` (list of {name, factor_type:"categorical", levels:[...]}) makes the
optimizer search over the levels of categorical factors instead of treating them as numbers.Inputs: data (required), response_specs (required), method, api_key, +2 more. Returns: result
covadoe_plot_dataAInspect
Generate plot data for DoE visualizations (14 plot types).
Analysis plots (need analysis_result, or design_data+response_col to auto-fit): main_effects,
interaction, pareto_effects, half_normal, normal_effects (signed Daniel plot; `alpha` sets the
reference threshold). Model plots (need model_result, or design_data+response_col): contour_2d,
surface_3d, residuals_4plot, prediction_profiler, cube_plot, ternary_mixture. Optimization plots:
desirability, pareto_front. `overlaid_contour` (sweet-spot / green-zone) takes `response_specs`
(list of {name, lower/upper}) + `factor_x`/`factor_y`; it fits one RSM per response from
design_data and returns the combined in-spec mask (`fixed_factors` pins the other factors).Inputs: plot_type (required), alpha, api_key, factors, +13 more. Returns: result
covadoe_post_hoc_powerCInspect
Post-hoc power analysis from existing experiment.
Inputs: data (required), response (required), api_key, model_terms, +1 more. Returns: result
covadoe_power_analysisAInspect
Statistical power ASSUMING AN ORTHOGONAL ±1 DESIGN — prefer covadoe_design_power.
Takes only scalars, so it never sees the design and cannot check its own premise. It uses
`ncp = n_runs · effect_size²`, which requires `(X'X)⁻¹_jj = 1/n_runs` — true for two-level
factorial and Plackett-Burman designs, and FALSE for a central composite, a Box-Behnken, a
definitive screening design, or a plain factorial WITH CENTRE POINTS.
On those it is wrong by a lot, and confidently: 0.8677 against a true 0.6228 on a Box-Behnken,
0.3874 against 0.2909 on a DSD, 0.3874 against 0.2491 on a 14-run factorial with centre points
(each Monte Carlo confirmed, 2026-08-07).
Use `covadoe_design_power` for ANY design carrying centre or axial points — it builds the design
and takes the non-centrality from `(X'X)⁻¹`, needs no data, and also reports what this tool would
have claimed. Use this one only when the design really is an orthogonal two-level design.
Every response carries `assumes_orthogonal_design: true` and a `validity` note.Inputs: n_runs (required), n_terms (required), effect_size (required), alpha, +2 more. Returns: result
covadoe_predictCInspect
Predict response at given factor settings with CI and PI.
Inputs: model (required), settings (required), alpha, api_key. Returns: result
covadoe_residual_diagnosticsCInspect
Residual diagnostics: normality (Shapiro-Wilk), Cook's distance, leverage, and layout_check: residuals vs run order (trend, Durbin-Watson) and plate position.
`factor_info` (list of {name, factor_type:"categorical", levels:[...]}) dummy-codes
categorical factors. `plate_layout` maps roles (well, row, column, plate, run_order,
internal_standard) to columns; a `run_order`/`RunOrder` column is used automatically.Inputs: data (required), response (required), api_key, factor_info, +3 more. Returns: result
covadoe_ridge_analysisAInspect
Ridge / steepest-ascent path of a quadratic RSM (Draper; Myers & Montgomery).
When the unconstrained optimum is a saddle or sits outside the studied region, this traces the
locus of best predicted response on spheres of increasing radius around the design center — the
direction to run the next experiments. Pass a fitted `model` (from fit_rsm) OR raw `data` +
`response` to auto-fit one. Returns the eigenvalue nature, stationary point, and the ridge `path`
(radius, real-unit settings, predicted response ± CI) ending at `recommended_settings`.Inputs: data, goal, model, api_key, +5 more. Returns: result
covadoe_suggest_designBInspect
Recommend a design type based on number of factors and goal.
Pass `n_categorical` (how many of the n_factors are categorical, e.g. solvent,
ligand, base): classical designs such as CCD cannot carry them.Inputs: n_factors (required), goal, api_key, n_categorical. Returns: result
covadoe_taguchiAInspect
Taguchi robust-design analysis: signal-to-noise ratios + main effects on the S/N.
Computes the S/N ratio per run and the factor-level S/N means to find the most robust setting.
`sn_type` selects the ratio ("larger_better"|"smaller_better"|"nominal_better"); if omitted it is
inferred from `response_goal` ("maximize"|"minimize"|"target"). Returns per-factor S/N tables,
the recommended robust level per factor, and a delta/rank contribution ordering.Inputs: data (required), response (required), api_key, sn_type, +2 more. Returns: result
covair_annotateCInspect
Annotate IR spectrum with known molecular structure (SMILES).
Inputs: path (required), smiles (required), api_key. Returns: result
covair_auto_analyzeCInspect
Full IR analysis in one call: load, optimize processing, detect peaks, report.
Inputs: path (required), smiles, api_key. Returns: result
covair_batchCInspect
Batch-process all IR spectra in a folder with auto-optimized parameters.
Inputs: folder (required), api_key. Returns: result
covair_compareCInspect
Compare two IR spectra: correlation, distance, verdict.
Inputs: path_a (required), path_b (required), api_key. Returns: result
covair_identifyCInspect
Identify unknown IR spectrum: functional groups, substance class, contaminants.
Inputs: path (required), api_key, library_path. Returns: result
covair_predict_irCInspect
Predict expected IR absorption bands from SMILES string.
Inputs: smiles (required), api_key. Returns: result
covair_processCInspect
IR analysis with explicit processing parameters (baseline + normalization).
Inputs: path (required), api_key, baseline, normalize_method. Returns: result
covair_quality_checkBInspect
Assess IR spectrum quality: S/N, water vapor, CO2, saturation, grade.
Inputs: path (required), api_key. Returns: result
covair_verifyAInspect
Verify a structure against a measured IR spectrum: verdict, band recall, and what is left unexplained.
Pass `path` whenever you have the spectrum file. The curve is baseline-corrected
and normalised the way covair_auto_analyze does it, and the verdict then requires
two things: that the structure's expected bands were found AND that the structure
accounts for the absorption actually measured. With neither a `path` nor peaks
there is nothing to verify against and the verdict is `no_evidence` -- not
`inconsistent`, which would assert evidence against the structure where none was
supplied. With peaks but no curve, `verdict_basis` reads `peak_recall_only` and
the verdict cannot reach `consistent`.
Do NOT read a verdict off `match_score`. It is the fraction of the CANDIDATE's own
expected bands that were found, so the denominator belongs to the candidate and a
structure predicting few bands scores high on any spectrum. Measured on a real
aspirin spectrum: hexane, which contains no oxygen, scores 1.00 where aspirin
scores 0.92, because all four of hexane's expected bands are C-H. Read `match_score`
with `n_expected_bands` beside it; `band_recall_verdict` shows what recall alone
would have concluded, and `verdict_note` says in words why the verdict differs.
Two measured limits. This blocks the "explains almost nothing" class, not wrong
structures in general -- nitrobenzene still reads `consistent` against an aspirin
spectrum. And on featureless input the unexplained-absorption test collapses onto
its chance level, so a broad candidate can pass on noise; check
`claimed_axis_fraction` in that case.
Inputs: smiles (required), path, api_key, tolerance_cm. Returns: result
covams_analyze_unknownCInspect
Full automated analysis of an unknown MS/MS spectrum (quality + processing + identify + classify + annotate).
Inputs: peaks (required), precursor_mz (required), adduct, api_key, +9 more. Returns: result
covams_compound_classCInspect
Predict compound class from MS/MS spectrum using diagnostic fragment rules.
Inputs: peaks (required), precursor_mz (required), api_key. Returns: result
covams_degradation_predictCInspect
Predict degradation products for an API molecule.
Inputs: api_smiles (required), api_key, pathways. Returns: result
covams_detect_featuresAInspect
Detect LC-MS features (m/z + retention time + intensity + charge) from a whole mzML/mzXML run.
Unlike the peak-list MS tools, this takes the RAW LC-MS FILE by reference
(path = a signed URL or local path) and returns EVERY detected feature, not
one collapsed spectrum — the entry point for untargeted metabolomics. Uses
pyOpenMS FeatureFinderMetabo when available, else a scipy m/z-binning
fallback (the response's method_used field reports which ran).Inputs: path (required), api_key, noise_threshold, min_trace_length, +1 more. Returns: result
covams_extractables_leachablesCInspect
Screen sample peaks against known extractable/leachable (E&L) compounds.
Inputs: sample_peaks (required), api_key, container_type. Returns: result
covams_forced_degradationCInspect
Run forced degradation study across multiple stress conditions (ICH Q1A/Q1B).
Inputs: api_smiles (required), stress_data (required), api_key. Returns: result
covams_formulaCInspect
Resolve molecular formula candidates from a measured m/z value.
Inputs: measured_mz (required), adduct, api_key, tolerance_ppm. Returns: result
covams_fragment_annotateCInspect
Annotate measured MS/MS peaks against predicted fragments for a structure.
Inputs: peaks (required), smiles (required), api_key, tolerance_ppm. Returns: result
covams_fragment_predictCInspect
Predict MS/MS fragmentation spectrum from a SMILES structure.
Inputs: smiles (required), adduct, api_key, collision_energy. Returns: result
covams_identifyCInspect
Identify an unknown compound from MS/MS peaks and precursor m/z.
Inputs: peaks (required), precursor_mz (required), adduct, api_key, +5 more. Returns: result
covams_identify_batchBInspect
Identify multiple spectra in parallel (batch mode for LC-MS/MS runs).
Inputs: spectra (required), api_key, databases, enable_sota, +2 more. Returns: result
covams_impurity_findCInspect
Find and identify impurities by matching against degradation products.
Inputs: api_smiles (required), unknown_peaks (required), api_key. Returns: result
covams_impurity_profilingCInspect
ICH Q3A/Q3B compliant impurity profiling with regulatory thresholds.
Inputs: api_smiles (required), batch_peaks (required), api_key. Returns: result
covams_isotopeCInspect
Predict theoretical isotope pattern for a molecular formula.
Inputs: formula (required), charge, api_key, n_peaks. Returns: result
covams_mass_balanceCInspect
Calculate mass balance from API and impurity peak areas.
Inputs: api_area (required), impurity_areas (required), api_key. Returns: result
covams_metabolite_idCInspect
Identify drug metabolites by mass shift matching (Phase I and Phase II).
Inputs: drug_smiles (required), metabolite_peaks (required), api_key. Returns: result
covams_neutral_lossCInspect
Annotate neutral losses between precursor and fragment peaks.
Inputs: precursor_mz (required), fragment_peaks (required), api_key. Returns: result
covams_predict_fingerprintAInspect
Predict a molecular fingerprint (ECFP4-style bit probabilities) from an MS/MS spectrum.
A CSI:FingerID-style structural descriptor of the unknown, inferred from its
fragmentation. Useful as a scaffold-level fingerprint to compare against
candidate structures (see covams_search_by_fingerprint).Inputs: peaks (required), precursor_mz (required), adduct, api_key, +1 more. Returns: result
covams_predict_rtCInspect
Predict chromatographic retention time for a molecule.
Inputs: smiles (required), api_key, model_path. Returns: result
covams_process_spectrumCInspect
Pre-process an MS spectrum (smoothing, peak picking, deisotoping, normalization).
Inputs: peaks (required), steps, params, api_key. Returns: result
covams_search_by_fingerprintAInspect
Shorten a candidate list for an UNKNOWN spectrum — triage, not identification.
Predicts the query's fingerprint from its peaks and ranks each candidate
SMILES by Tanimoto agreement. The ranking rests only on the subset of bits
the predictor actually learned; the rest sit at an untrained 0.5 prior and
would otherwise order candidates by molecular size. That subset is a small
fraction of the fingerprint, so `bit_coverage` in the response is the
authoritative live number — read it, and the `warning` that accompanies a
low one, before acting on the order.
Measured on open structure retrieval against MassSpecGym candidate pools
(median 256 candidates): the correct structure comes first about 3.5 % of
the time, against a 2.4 % permutation noise floor. The lift is real but
roughly one percentage point, so the right answer is NOT at rank 1 in about
96 % of cases. Use it to cut a long candidate list down, never to name a
compound.
If you already suspect what the compound is, use covams_search_db instead:
matching the spectrum against the packaged 169k-spectrum library is a far
stronger signal, and dereplicating a known compound is the common case.Inputs: peaks (required), precursor_mz (required), candidate_smiles (required), top_k, +2 more. Returns: result
covams_search_dbCInspect
Search spectral databases for matching compounds by spectral similarity.
Inputs: peaks (required), precursor_mz (required), api_key, databases, +3 more. Returns: result
covams_spectral_qualityCInspect
Assess quality of an MS/MS spectrum (grade A through D).
Inputs: peaks (required), api_key, precursor_mz. Returns: result
covasolve_predictAInspect
SOLUBILITY PREDICTION — gradient-boosted model (ensemble_v12) trained on 115,862 experimental solubility rows (BigSolDB and other sources).
**Always use this tool when the user asks about solubility, log S, dissolution,
or "how much X dissolves in Y".** Much more accurate than the heuristic logS
in `covachem_profile` and explicitly accounts for the solvent (covachem only
gives intrinsic water-logS).
Returns: log S (mol/L), mg/mL, g/L, mole fraction (dilute approximation with the
solvent's real density; null with a warning when the solvent's density is not in
the solvent table), applicability-domain verdict, warnings, and model_version.
Uncertainty: use `conformal_interval_90` — a calibrated 90 % interval on log S
(measured coverage about 0.90 on held-out data). With v12, `confidence` and
`confidence_interval_95` are null: the model's seed replicates do not give a
meaningful spread, so no such number is reported.
Solvent SMILES examples:
water = "O"
ethanol = "CCO"
methanol = "CO"
DMSO = "CS(C)=O"
acetone = "CC(C)=O"
ethyl acetate = "CCOC(C)=O"
Cost: 5 credits, available on Free tier.Inputs: solute_smiles (required), api_key, temperature_k, solvent_smiles. Returns: result
covasyn_cancel_jobAInspect
Stop a running async GPU job and refund its credits.
Parameters:
job_id: The id returned by the submitting tool, e.g. "covafold-fold_protein:abc123".
Returns: {status: cancelled|already_finished|not_found, message}.
FREE — and the job's own credits are returned automatically.
Inputs: job_id (required), api_key. Returns: result
covasyn_creditsAInspect
Show your current credit balance, tier, and pre-paid pack pricing.
Returns:
- balance: subscription credits remaining (resets weekly to weekly_allocation, capped at max_balance)
- prepaid_balance: pre-paid credits remaining (no expiration; consumed after sub balance is empty)
- total_available: balance + prepaid_balance (the actual amount you can spend right now)
- weekly_allocation: sub credits added every 7 days
- max_balance: rollover ceiling for sub credits (4× weekly_allocation)
- tier: free / pro / unlimited / enterprise
- max_concurrent_calls: parallel call limit for your tier
- next_reset_at: ISO timestamp when next weekly grant happens
- pricing_version: currently active pricing version
- prepaid_packs: list of available credit-pack sizes + EUR prices for your tier
This tool is FREE in all tiers — never costs credits to check your balance.
Inputs: api_key. Returns: result
covasyn_feedbackAInspect
Submit feedback to the CovaSyn team.
Use this when something doesn't work, you have a feature request,
or a question. Your data (SMILES, structures, results) is NEVER
included — only your message and optional contact info.
Args:
subject: Short summary of the feedback.
message: Detailed description.
severity: One of "bug", "feature", "question", "other".
contact: Optional email/name so we can follow up.
tool_name: Which tool the feedback is about (optional).
error_message: Error text if reporting a bug (optional).
api_key: Your API key (used for rate limiting only).
Inputs: message (required), subject (required), api_key, contact, +3 more. Returns: result
covasyn_job_detailAInspect
Show one job's inputs — the sequence it was given, and its timings.
Parameters:
job_id: The id from covasyn_jobs, e.g. "covafold-fold_protein:abc123".
Returns: {job: {...}} or {error}. FREE — never costs credits.
Inputs: job_id (required), api_key. Returns: result
covasyn_job_resultAInspect
Fetch the result of an async GPU job (returned by a long-running tool).
Some GPU tools return a ``job_id`` instead of an inline result because the
computation runs too long for one response (e.g. chemreason_retrosynthesis).
Call this with that job_id; if it is not ready, re-call every ~20 seconds
until ``status`` is ``completed`` or ``failed``.
Parameters:
job_id: The id returned by the submitting tool, e.g. "chemreason:abc123".
Returns: {status: running|completed|failed, result?, error?}.
FREE — never costs credits.
Inputs: job_id (required), api_key. Returns: result
covasyn_jobsAInspect
List your async GPU jobs — running, finished and failed.
Parameters:
scope: "canvas" (default) for jobs started from the current canvas,
or "account" for every job on your account.
tool: optional filter, e.g. "covafold". Empty (default) returns EVERY
durable job — an LLM asked "what jobs do I have?" must not be
handed a silently truncated answer. A UI that can only act on one
kind of job passes its own filter explicitly.
Returns: {jobs: [{job_id, tool, state, submitted_at, credits, importable}]}.
FREE — never costs credits. Never returns sequences or structures; fetch
those with covasyn_job_detail / covasyn_job_result.
Inputs: tool, scope, api_key. Returns: result
covatox_ecotoxAInspect
Environmental fate screen: bioaccumulation (BCF), persistence (biodegradation half-life, ready biodegradability), ecotox structural alerts, and the REACH P/B criteria.
Does NOT report acute aquatic toxicity. The fish 96 h LC50, Daphnia 48 h
EC50 and algae 72 h ErC50 come back as "not_served" with a reason: on
held-out EPA ECOTOX reference values these predictors show no usable
ranking skill and a larger error than a single fixed number, so no value
is published that a user could not tell apart from a real one. For acute
aquatic toxicity use a measured OECD 203/202/201 result or the EPA ECOTOX
Knowledgebase. Consequently no GHS aquatic category is returned, and the
PBT T criterion is reported as unassessed rather than negative.Inputs: smiles (required), api_key. Returns: result
covatox_healthAInspect
Liveness/readiness check + cache occupancy. Returns status:
ready (warm), warm_required (models on disk but caches empty),
or no_models. Use as K8s liveness/readiness probe.
Inputs: api_key, model_dir. Returns: result
covatox_warmupAInspect
Pre-load all CovaTox inference caches (models, AD references, conformal calibrators). Call once after container start so the first user prediction skips the ~5-10s of cold disk reads. Idempotent.
Inputs: api_key, model_dir. Returns: result
covauv_auto_analyzeBInspect
Full UV-Vis spectrum analysis in one call (recommended entry point).
Loads, processes, detects peaks, identifies chromophores, assesses quality, and formats UV-Vis reporting strings.
Inputs: path (required), smiles, api_key, solvent. Returns: result
covauv_bioCInspect
Biomolecule analytics: protein A280, DNA/RNA A260, or enzyme kinetics.
mode="protein" | "nucleic_acid" | "enzyme_kinetics"
Inputs: mw, a260, a280, mode, +7 more. Returns: result
covauv_compareBInspect
Compare two UV-Vis spectra for similarity (correlation/euclidean/cosine).
Inputs: path_a (required), path_b (required), method, api_key. Returns: result
covauv_dissolutionAInspect
Calculate dissolution profile from UV-Vis absorbance time series.
concentration_unit is REQUIRED and must match the unit the calibration
curve was built in: "mg/mL", "mg/L" or "ug/mL" (the latter two are
equivalent). Dissolution calibrations are most often built in ug/mL — a
100 mg tablet in 900 mL is ~111 ug/mL at full release. Getting this wrong
shifts every result by a factor of 1000, so there is deliberately no
default; the response echoes the unit back and flags implausible profiles
in unit_check.
Optionally calculates f2 similarity if reference_profile is provided.
Inputs: timepoints (required), absorbances (required), label_claim_mg (required), calibration_slope (required), +5 more. Returns: result
covauv_predictCInspect
Predict UV-Vis absorption bands and phototoxicity risk from SMILES.
Inputs: smiles (required), api_key, solvent. Returns: result
covauv_quantifyAInspect
Quantification via Beer-Lambert law or calibration curve.
Mode 1: Provide 2 of (absorbance, epsilon, concentration) for Beer-Lambert.
Mode 2: Provide standards JSON + unknown_absorbance for calibration.
Inputs: mw, api_key, epsilon, standards, +4 more. Returns: result
covauv_validateCInspect
Run full ICH Q2(R2) method validation (linearity + accuracy + precision).
Inputs: standards (required), accuracy_data (required), precision_data (required), api_key. Returns: result
covnmr_2d_analyzeAInspect
Analyze one 2D NMR experiment (HSQC/HMBC/COSY/NOESY/TOCSY/ROESY): crosspeaks and the correlations they imply.
The experiment is read from the file's own pulse sequence; pass
`experiment` only for formats that record none. A value contradicting the
file is reported in `experiment_warning` and does not override it. To use
2D data for an atom-by-atom assignment, give the files to covnmr_sudoku.
Inputs: path (required), api_key, experiment, max_crosspeaks. Returns: result
covnmr_analyzeAInspect
Full 1D NMR analysis of one spectrum file: processing, peaks, integrals,
multiplets, the publication NMR string, and, with smiles, structure
verification in the same call.
With `smiles`, pass `hrms` too (e.g. {"formula": "C9H8O4", "mz": 181.0495,
"adduct": "[M+H]+"}) and `verification` checks the HRMS formula and mass
(within `hrms_ppm_tolerance` ppm, default 10), the proton count from the
spectrum's integrals (1H) or the carbon line count (13C), and gives one
answer in `verification.verdict_combined` (see covnmr_verify).
Absorbs three retired tools:
* covnmr_format -> `output_format="nmr_string"` (or "acs").
* covnmr_export -> `output_format` "peaks_csv" (alias "csv"), "spectrum_csv",
"jcamp" or "nmredata" (pass `smiles` and `compound_name` for NMReData).
The file comes back INLINE as `content` with `filename` and `mime`;
nothing is written on the server. `content_truncated` says whether it was
capped at `max_content_chars` (2,000,000); curve formats are decimated to
fit rather than cut (`curve_reduced_to_fit`).
* covnmr_audit_report -> `audit=true` (and `operator`): adds `audit_record`
with input and output hashes taken over the complete output. The record
names your file: `display_name` if you give one, else the file name in
the URL you sent.
`output_format="asm"` returns an Allotrope-style ASM JSON document inline.
It is EXPERIMENTAL and UNVALIDATED: no NMR ASM schema was available to
check it against, and the response says so (`maturity: "experimental"`,
`maturity_note`, `asm_validation`). Do not hand it to a system that
expects a validated ASM file.
Nucleus and spectrometer frequency are read from the file; give them only
for a file that declares none. Without `smiles` the 1H entries of the NMR
string carry no integrals (there is no honest anchor to scale them
against); with it they do. With `smiles` the response also adds
`verification` and `impurities` -- read `verification.verdict` and its
`evidence`, not `match_score`. For a structure check alone, or one from a
peak list without a file, use covnmr_verify.
Inputs: path (required), hrms, audit, smiles, +11 more. Returns: result
covnmr_batchAInspect
Batch-analyze every NMR data file in an archive (zip/tar URL): per-file
summaries plus statistics, files_processed and files_failed.
Nucleus and spectrometer frequency are read from each file, so an archive
holding spectra from two instruments is handled correctly and every row
reports its own `field_mhz` and `field_source`. Give `field_mhz` only for
files that declare no frequency; a file that states one always wins. An
archive without NMR data returns status "error" (`no_data_found`).
Inputs: folder (required), api_key, nucleus, solvent, +1 more. Returns: result
covnmr_compareAInspect
Compare two NMR spectra of the same nucleus: matched peaks, correlation, RMS shift difference and a similarity verdict (e.g. batch against reference).
Each file declares its own nucleus; two different nuclei return status
"error" (`nucleus_mismatch`, with `nucleus_a` and `nucleus_b`) rather than
a meaningless score. Peaks are matched within `tolerance_ppm` (0.05).
Inputs: path_a (required), path_b (required), api_key, nucleus, +2 more. Returns: result
covnmr_explainAInspect
Interpret ONE peak, or one 2D crosspeak: what is this signal? The answer
is in explanation.
Peak mode (default): `ppm` and optionally `multiplicity` ("t", "dd"),
`j_values` (comma-separated Hz, "7.1,2.4") and `nucleus` (default 1H).
Returns the spectral region, the functional groups consistent with it, and
what the coupling pattern implies.
Correlation mode: give `c_ppm` AND `experiment` ("HSQC", "HMBC", "COSY")
and it interprets the crosspeak instead -- bond type and the groups on each
axis. This mode returns a different key set; `mode` says which you got.
With `smiles` you also get `atom_attribution`: the atom whose predicted
shift is closest to `ppm`. It is a nearest neighbour on one axis, not an
assignment, and it names an atom even when nothing is within tolerance --
read `attribution_status` first ("matched", "no_atom_within_tolerance",
"no_structure_prediction", "not_requested"). For a full assignment use
covnmr_verify or covnmr_sudoku.
Inputs: ppm (required), c_ppm, smiles, api_key, +5 more. Returns: result
covnmr_identifyAInspect
Identify a spectrum against a curated reference set: 7 qNMR standards + 32 common impurities/solvents (39 entries total) -- not a general or natural-product compound database. Good for confirming a known standard or a common impurity; it will not identify an unknown natural product or a novel structure.
Nucleus coverage differs and the response states it in `nucleus_coverage`.
1H: all 39 entries, 7 solvents. 13C: 30 of the 32 impurities, CDCl3 and
DMSO-d6 only (Fulmer et al., Organometallics 2010, 29, 2176-2179, on the
77.16 scale); the 7 qNMR standards carry no 13C. A nucleus with no
reference data returns status "error" (`unsupported_nucleus`).
Read the top score, not the ranking: a random 8-peak carbon list averages
0.24 against its best entry and has reached 0.625, while every compound
that IS in the table self-matches at 0.800 or better. Only a score near 1.0
is an identification; anything lower is the nearest neighbour in a
39-compound table. To test a structure you already suspect, use
covnmr_verify; to choose between suspects, covnmr_rank_candidates.
Inputs: path (required), api_key, nucleus, solvent, +2 more. Returns: result
covnmr_impurityAInspect
Impurity profiling + ICH Q3C(R8) residual-solvent screen of a spectrum.
The NMR solvent the file declares wins over `solvent`. Residual solvents
are named by chemical shift for lines the compound (`smiles`) does not
explain; the screen does not quantify them, so it gives limits, not
pass/fail. Each identified line lists every solvent within tolerance
(`ambiguous`, `alternatives`), its ICH class, and `limit_ppm_at_dose`.
Without `dose_g_per_day` the Option 1 dose of 10 g/day is used and
`dose_source` says so. `coverage` names the Q3C solvents this NMR solvent
cannot screen for, so a clean result can be read correctly. A nucleus with
no tables returns status "error" (`unsupported_nucleus`).
Inputs: path (required), smiles, api_key, nucleus, +2 more. Returns: result
covnmr_method_validateAInspect
ICH Q2(R2) method validation: linearity, precision, LOD/LOQ.
`concentrations` and `responses` are comma-separated series of equal
length, AT LEAST FIVE points ("1,2,5,10,20"). Linearity is tested with
Mandel's F-test, so it is a verdict on curvature, not an R-squared
threshold. `overall_passes` is true only when every criterion holds:
positive slope, r2 >= 0.999, Mandel linear, and RSD <= 1 % when
`precision_measurements` (comma-separated repeats) are given; the criteria
are listed in `criteria`.
Give `noise_rms` for LOD/LOQ from the calibration slope; without it
`lod_loq.status` is "not_computed" with the reason. Bounded: `max_points`
(1000) per series, `max_residuals` (500) on the residuals echoed back.
Inputs: responses (required), concentrations (required), api_key, noise_rms, +3 more. Returns: result
covnmr_parse_nmr_stringAInspect
Publication-style NMR text -> a structured peak list.
Paste the string as it appears in an SI or a paper, e.g.
"13C NMR (100 MHz, CDCl3): d 58.4, 18.2" or
"1H NMR (400 MHz, CDCl3): d 7.26 (d, J = 8.1 Hz, 2H), 1.25 (t, 3H)".
The nucleus is read off the header and stamped on every peak, which is
what stops a pasted 13C list from being scored against proton
predictions; text with several headers comes back in `peaks_by_nucleus`.
Multi-step: pass `peaks` to covnmr_verify (one structure) or
covnmr_rank_candidates (several). Bounded at `max_peaks` (1000);
`n_parsed` is the untruncated count. Read `warnings`: every value that was
repaired (e.g. "149. 6" read as 149.6) or dropped is named there, so a
clean parse has `warnings: []`.
Inputs: text (required), api_key, max_peaks. Returns: result
covnmr_predictAInspect
Predict NMR from a SMILES, no spectrum needed: chemical shifts, a synthetic spectrum, or a DEPT subspectrum. For planning and comparison before measurement.
`output`:
* "shifts" (default): per-atom predicted shifts for `nucleus` (default
1H). Where an atom has an uncertainty interval it is given, and where it
has none the response says why; `predictor` says which predictor
answered and for how many atoms.
* "spectrum": the coupling-resolved synthetic spectrum with multiplets,
J-couplings and integrals. Absorbs the retired covnmr_predict_spectrum.
* "dept135" / "dept90": a 13C DEPT subspectrum (CH/CH3 positive, CH2
negative, quaternaries absent; DEPT-90 CH only); each peak carries its
`sign` and `dept_class`. Absorbs the retired
covnmr_predict_dept_spectrum.
The curve is decimated to `max_points` (2048, ceiling 16384; 0 returns the
peak list only). Without `field_mhz` a 400 MHz spectrometer is assumed and
`field_mhz_source` says "default". To score the prediction against a
measured spectrum, use covnmr_verify.
Nuclei: 1H, 13C and 19F. 31P is REFUSED (`unsupported_nucleus`): there is
no 31P predictor, only one typical value per bonding class, and it missed
real compounds by 21-38 ppm. Pass `allow_experimental=true` to get it
anyway, marked `maturity: "experimental"` with a +/-40 ppm uncertainty.
Inputs: smiles (required), output, api_key, nucleus, +5 more. Returns: result
covnmr_purityAInspect
qNMR purity with an internal standard (USP <761>), with an uncertainty
budget. Needs standard_protons, analyte_protons, analyte_mw (or
analyte_formula), and standard_conc_mm unless the masses are given.
NAME THE TWO SIGNALS: `standard_peak_ppm` (older name `reference_peak_ppm`)
and `analyte_peak_ppm`. A named signal that is not found within
`selection_tolerance_ppm` (0.3) is an error, never the nearest line
instead; 0.0 is a real shift (TMS). The response's `selection` block names
the regions integrated, so the number can be checked against the spectrum.
Optional `analyte_signals`, e.g. [{"ppm": 3.80, "protons": 3}, {"ppm": 1.20,
"protons": 6}], quantifies the analyte on each listed signal and reports
the signals that disagree beyond the integration uncertainty -- the usual
sign of an overlapping impurity. Give `sample_mass_mg` and
`sample_volume_ml` for a mass-based purity. `smiles` is accepted for older
callers and not used.
Traceable purity (USP <761> budget): give `analyte_mass_mg`,
`standard_mass_mg` and `standard_purity`. `purity_traceable` then holds the
mass-fraction purity, a row per input in `budget` (value, uncertainty,
share of the variance), the dominant source, and the expanded uncertainty
with its coverage factor. Inputs you gave no uncertainty for are filled
with stated defaults and listed in `assumed`. Name the standard with
`internal_standard` (e.g. "maleic_acid", "dss", "tms") to take its molar
mass and signals from the table; an unknown name is an error
(`unknown_internal_standard`), and a standard whose signals overlap an
analyte region is refused (`internal_standard_overlap`).
`system_suitability` always checks S/N, baseline, phase and relaxation;
give `t1_longest_s` and `relaxation_delay_s` (else read from the file) or
the relaxation test stays untested. A failed check is listed; it does not
withhold the number. `audit=true` adds `audit_record`, a hash-chained
record of inputs, steps and result (no electronic signature); it names
your file as `display_name`, else the file name in the URL. For ERETIC,
PULCON or a purity from integrals you already have, use
covnmr_qnmr_advanced.
Inputs: path (required), audit, smiles, api_key, +39 more. Returns: result
covnmr_qnmr_advancedAInspect
qNMR from integrals you already have: ERETIC, PULCON, a full uncertainty budget, or a traceable purity. For an internal-standard purity from a spectrum file use covnmr_purity.
`method` selects which arguments are read; a missing required one is an
error that names it (`missing`), never a silent default:
* "eretic" (electronic reference): `analyte_integral`, `eretic_integral`,
`eretic_response_factor`, `analyte_protons`, `analyte_mw`. Give
`sample_mass_mg` for `weight_percent`; without it `weight_percent` is
null and `weight_percent_status` says "not_computed".
* "pulcon" (reference in a separate tube): `analyte_integral`,
`analyte_protons`, `analyte_mw`, `reference_integral`,
`reference_protons`, `reference_conc_mm`; optional T1s, recycle delay and
the two fields. Check `t1_correction_factor` -- that is where an
over-short recycle delay shows up.
* "uncertainty" (GUM budget of an internal-standard result):
`analyte_integral`, `standard_integral`, `standard_conc_mm`,
`standard_protons`, `analyte_protons`, `analyte_mw`; optional
`standard_purity` and the u_* inputs. Returns `components`,
`dominant_source`, `combined_u_rel` and the k=2 expanded uncertainty;
`n_sources_default_constant` counts the sources you did not supply.
* "purity" (USP <761> mass-fraction purity, no spectrum):
`analyte_mass_mg`, `standard_mass_mg`, `standard_purity`,
`analyte_protons`, `standard_protons`; the ratio from
`analyte_integral` + `standard_integral`, `integral_ratio` or
`integral_ratio_replicates`; the molar masses from `analyte_mw` /
`analyte_formula` and `standard_mw` / `standard_formula` /
`internal_standard`. Returns the same budget as covnmr_purity's
`purity_traceable`, at top level, plus `system_suitability` (from your
`sn_standard`, `sn_analyte`, `relaxation_delay_s`, `t1_longest_s`) and
`internal_standard_check` (from the shifts you state; an overlap is the
error `internal_standard_overlap`). `audit=true` adds `audit_record`.
Each method returns a different key set; `method` is echoed back.
Inputs: method (required), audit, api_key, solvent, +57 more. Returns: result
covnmr_rank_candidatesAInspect
Rank several candidate structures against one measured spectrum.
Candidates: `candidates` (list of SMILES), `candidates_csv`
("CCO,CC(=O)O") or `candidates_json`. Spectrum: `path`, else a peak list
(`peaks` / `peaks_json`, `experiments` / `experiments_json`).
GIVE `path` IF YOU HAVE THE FILE: only the file makes the second axis --
overlap of each candidate's full predicted curve with the measured one --
available. Measured on a real 13C APT of a triterpene, the correct
structure ranked first on both axes against n-hexadecane and glucose
(curve 0.37 vs 0.06/0.08). With `path` the order is the curve axis and each
row says whether the peak axis agrees (`agreement`); with a peak list only,
the order is the evidence log-likelihood ratio. The peak axis DECLINES --
`ranking_basis` "declined_unassessable_peak_list", every `rank` null --
when a candidate's evidence cannot be computed (19F/31P, fallback
predictor). Read `ranking_basis` before `ranked`: a null `rank` is the tool
declining to answer, not a tie.
Capped at `max_candidates` (50). With `path` the file decides nucleus,
solvent and field (`nucleus_source`). Multi-step: confirm the winner with
covnmr_verify, or covnmr_sudoku when you have 2D data.
Inputs: path, peaks, api_key, nucleus, +10 more. Returns: result
covnmr_referenceAInspect
Look up NMR reference data -- no spectrum needed.
`category`:
* "shift_range" -- typical shift window for a functional group, for
`nucleus` (1H by default, or 13C). A nucleus with no table returns status
"error" (`unsupported_nucleus`) with `results: []`, never the proton
table.
* "coupling" -- J-coupling ranges by type (vicinal, geminal, ...).
* "solvent" -- residual solvent peaks, one record per solvent with its
1H / 13C / 19F / 31P shifts.
* "all" (default) -- every table, grouped matches.
An empty `query` is a legitimate "show me the table" request. Bounded at
`max_results` (50) per category; `n_results` is the untruncated count.
Inputs: query, api_key, nucleus, category, +1 more. Returns: result
covnmr_sstAInspect
USP <761> system suitability: linewidth (FWHM) of the reference line,
signal-to-noise, peak asymmetry. Also the quick quality check: it absorbs
the retired covnmr_quality, whose S/N block is now the quality field.
`verdict` is pass / fail / incomplete, or "reference_not_found" when no
line is in reach of `reference_peak_ppm` (default 0.0, TMS). S/N is
evaluated only when `signal_ppm` is given, and a suite with an unevaluated
criterion reports "incomplete" (`not_tested`), never a pass. Limits default
to `max_linewidth_hz` 1.0 and `min_sn` 250. The linewidth is the
instrument's (no line broadening unless `line_broadening_hz` is given); the
spectrometer frequency comes from the file unless `field_mhz` is given.
Inputs: path (required), min_sn, api_key, nucleus, +6 more. Returns: result
covnmr_sudokuAInspect
Assign every atom of a KNOWN molecule from several experiments at once (1H, 13C, APT, DEPT-135, DEPT-90, HSQC, HMBC, COSY), by constraint propagation across all of them.
Absorbs the retired covnmr_multi_analyze (`output="table"`: one row per
atom, per-experiment summaries, publication NMR strings, and the atoms
nothing could be assigned to with a reason) and the 2D form of
covnmr_assign (`paths_2d`: [{"experiment": "HSQC"|"HMBC"|"COSY",
"path": "<url>"}]). Default `output="assignments"` returns the solver's
assignment with `explanation`, which says why atoms stayed unassigned.
The carbon-edited slots say different things: an APT shows every carbon
and separates them by the PARITY of the proton count; a DEPT-135 shows only
protonated carbons (CH2 negative); a DEPT-90 shows CH only. Supply a
DEPT-135 together with a plain 13C whenever you have both: the quaternary
carbons come from the difference. The file decides which sequence it is --
a spectrum in the wrong slot is analysed as what it is and named in
`experiment_warnings`, as is every experiment that did not enter the
solve. `path` takes one 1D file whose own nucleus picks the slot.
Omit `tolerance_ppm` and each experiment gets its nucleus' window (1H 0.3
ppm, 13C 3.0). In the table, `sudoku_result.verdict` is the most cautious
of the per-nucleus verdicts; `match_score` is coverage and decides nothing.
Inputs: smiles (required), path, output, api_key, +15 more. Returns: result
covnmr_verifyAInspect
Does this NMR data match this structure (smiles, required)? Returns a
verdict, the evidence it was read from, and per-atom assignments.
The data is ONE of: a spectrum file (`path`), a peak list (`peaks`, or its
JSON text `peaks_json`), or peak lists keyed by experiment (`experiments` /
`experiments_json`, e.g. {"1H": [...], "13C": [...]}). Absorbs the retired
covnmr_verify_peaks (the peak-list form) and the 1D form of covnmr_assign.
Multi-step: paste literature text into covnmr_parse_nmr_string first and
pass its `peaks` here.
READ `verdict` AND `evidence`, NOT `match_score`. `verdict` ("match" /
"possible match" / "inconclusive" / "no match") comes from decoy-calibrated
evidence (`evidence.z`, the share of the spectrum left unexplained, and
thresholds set on decoy structures in `evidence.verdict_calibration`).
`match_score` is coverage (`match_score_basis`) and decides nothing. A
structure with fewer predicted signals than any calibrated one gets
"inconclusive"; 19F and 31P have no error model, so on their own they are
always "inconclusive".
A peak list that declares its nucleus outranks `nucleus`; a list spanning
several nuclei is scored per nucleus (`per_nucleus`) and the combined
verdict is the most cautious of the nuclei that have an error model. 19F
and 31P are still scored but do not set that verdict
(`nuclei_not_in_verdict` says why). With `path` the file decides the
nucleus.
`peaks_json` may also be the whole answer of covnmr_parse_nmr_string
({"nucleus": ..., "peaks": [...]}); its nucleus counts as the list's own.
Combine 1H/13C with HRMS, from a file or a peak list: pass `hrms`, e.g.
{"formula": "C9H8O4", "mz": 181.0495, "adduct": "[M+H]+"}. The response
then lists each check in `evidence_checks` -- HRMS formula, HRMS mass
within `hrms_ppm_tolerance` ppm (default 10), the 1H integral sum against
the proton count, the 13C line count against the carbon count, and the
shifts -- and gives one answer in `verdict_combined`: a failed HRMS check
makes it "no match", a failed count check caps it at "inconclusive",
otherwise it is the shift verdict. `verdict` itself is unchanged. The count
checks run on a file too (from its integrals and picked lines).
Give `candidates` (up to 20 alternative SMILES) and `verdict_relative` says
whether this structure beats the best alternative. It may also say the
data cannot tell this structure from a regioisomer: then it is
"inconclusive" with `regioisomer_ambiguous: true`, and
`competing_candidate` names the regioisomer that is not ruled out. Omit
`tolerance_ppm` and the window follows the nucleus (1H 0.3 ppm, 13C 3.0).
For several spectra or 2D data of one compound use covnmr_sudoku.
Inputs: hrms, path, peaks, smiles, +11 more. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Turn any LLM into your lab assistant: search samples, track experiments, analyze data with AI.
Find open biology and chemistry AI models, bioinformatics tools and databases, and run them online.
Biotech rNPV/PoS engine for AI agents. Signed exports, evidence register, asset landscape.
FTIR spectral search and material identification for AI agents via MCP.
Related MCP Servers
- AlicenseAqualityBmaintenanceEnables AI agents to call deterministic, cited, and tested implementations of ICH/FDA guidelines for pharmaceutical impurity assessment, stability and shelf-life evaluation, and SDTM conformance validation.17Apache 2.0
- AlicenseBqualityDmaintenanceEnables molecular design and simulation through 45 chemistry tools including pKa calculations, geometry optimization, conformer searches, docking, protein cofolding, and ADMET predictions powered by Rowan's computational chemistry platform.4124MIT
- AlicenseCqualityFmaintenanceEnables LLM agents to draw and edit ChemDraw reaction schemes, parse ELN exports and LCMS/NMR data, and generate publication-ready CDXML files through grounded chemistry tools.353MIT
- AlicenseNot gradedqualityBmaintenanceChemical safety intelligence for AI-assisted experiment design. 18 MCP tools for compatibility checks, hazard analysis, PPE, storage, waste disposal, mixing order, exposure limits, transport classification, regulatory compliance (8 jurisdictions), and signed audit reports. Supports 4,200+ chemicals in 5 languages.MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.