Igniva™ for Chemistry

A modular suite of AI agents for computational chemistry tasks – from structural biology through molecule and material discovery to manufacturing and controls, built on physics based methods such as quantum mechanics and molecular dynamics as well as generative AI models.

01

Drug discovery

  • Target and pocket elucidation
  • Hit generation and screening

  • Hit-to-lead and lead optimization

02

Molecular profiling

  • Developability and degradability

  • pKₐ and solubility

  • Spectra prediction

03

Drug substance

  • Solid-form and polymorph screening

  • Reaction and catalysis modelling

04

Drug product

  • Excipient selection and compatibility

  • Amorphous solid dispersion stability 
R&D WORKFLOWS

Applications & Agents

01

Drug discovery

From a prepared structure and a pocket, through screening, to the poses and analogues that carry a lead forward.

Target and pocket elucidation

Simulation-ready structures and the conformer ensembles every later calculation starts from.


HAddiT

Hydrogen addition with bond orders and valency handled correctly, so a structure is simulation-ready before any calculation runs on it.


EAGLE™

Environment-aware conformer ensembles for the setting that matters — crystal packing, a given solvent or a protein pocket — including macrocycles and cyclic peptides.


MolConSUL

Conformer generation using unsupervised learning.

Hit generation and screening

Finding, filtering and generating the molecules worth a closer look.


ScreenXcel

GPU-accelerated virtual screening.


MolScreener

Medicinal-chemistry filtering of candidate molecules.


MolMatch

Analogue search across ultra-large molecular libraries.


MolGenAI

Generative de novo molecular design.

Hit-to-lead and lead optimization

The binding pose an affinity estimate depends on, and the analogues worth making next.


ChemTwin

Local chemical-space exploration around an input molecule.


PosePRO

Binding-pose prediction for flexible pockets: induced-fit docking with interfacial water and perturbation MD, for on-target, off-target and anti-target proteins.

02

Molecular profiling

The properties that decide whether a molecule can be developed, predicted per site, per solvent or per atom.

Developability and degradability

Whether a molecule gets where it must, and what it turns into on the way.


NitroSafeAI

Nitrosation risk screening for APIs and excipients, scoring each nitrogen atom in the structure and benchmarked on public data.


CycPepPer

Cyclic-peptide permeability classification with counterfactual design.

pKa and solubility

Which site ionises, and which solvent to use, per solvent rather than per molecule.


SitepKa

Site-specific pKa prediction in water and 34 organic solvents, giving the site that ionises rather than a single value for the molecule.


SolviScope

Solubility and solvent ranking for solvent and antisolvent selection, with the report stating where the method applies and where it does not.

03

Drug substance

The solid the API is made as: which packings are plausible, and what they weigh.

Solid-form and polymorph screening

Crystal structures cleaned up, ranked and measured before a form is chosen.


CrystalClean Pro

Automated crystal-structure cleanup and supercell generation.


LENet

Machine-learned lattice-energy ranking for crystal screening.


TruRho

True-density estimation for active pharmaceutical ingredients.

04

Drug product

The formulation around the API: where to start, and which excipients behave.

Excipient selection and compatibility

Formulation starting points for a new API, and how its excipients behave in solution.


API surrogate search

A searchable portal of approved APIs with their physicochemical data and the inactive ingredients of approved formulations, for formulation starting points.


MicelleAI

Critical micelle concentration for surfactants and mixtures.

FAQ

Deployment

In your own cloud VPC or HPC, or as SaaS

Models

With your preferred LLMs and agentic AI platform

Data

Stays inside your boundary in VPC and on-premises deployments No training on proprietary data by default

Audit

Role-based access, with audit logs for every run and export

Agent packs by analysis type

Igniva™ runs agents from these packs contextually with a scientist signing off at key checkpoints. 

MOLECULAR DYNAMICS

How does this molecular system behave over time?

CRYSTAL PREDICTION

Which crystal packings should we investigate?

SMALL MOLECULE DISCOVERY

Which molecules should move forward for investigation?

FORMULATION

Which drug–excipient combinations deserve closer study?

DATA CURATION

How can papers and patents become usable scientific data?

RESULTS

Select Case Studies

Small molecule · Physico-chemical properties

Site-specific pKa in water and 34 organic solvents

SitepKa output for one molecule in DMSO: A predicted pKa for each ionizable site. 2 of its 3 ionizable sites are shown.

Methods

Dual-graph neural network · solute–solvent cross-attention · Kamlet–Taft solvent descriptors · DFT


Challenge

Process chemists work in organic solvents, but most deep-learning pKa models predict only in water, or give one value per molecule instead of one per site. Physics-based methods handle other solvents but are slow for large molecules.


Approach

SitepKa predicts a pKa for every ionizable site. Trained on 33,269 experimental values from ~10 public sources mined using Aganitha’s data curation agents.


Evidence

Preprint: SitepKa, ChemRxiv, May 2026, presented at ACS Fall 2026. Mean absolute error is 1.49–2.30 on the SAMPL6–8 aqueous blind challenges. For large molecules in 9 non-aqueous solvents, a DFT protocol with Aganitha’s in-house solvent-specific conformer generation pipeline, achieved an MAE of 2.23 on 52 validation molecules, ~100× faster than gold-standard physics protocols.

Outcome: pKa for each ionizable site across solvents used widely in the industry, to guide salt, co-crystal and crystallization-solvent choices

Impurity control · Regulatory CMC · Toxicity

NitroSafeAI: Nitrosation risk screening for APIs and excipients

NitroSafeAI’s knowledge-based & Deep-learning based scoring of ranitidine: a score and likelihood for each nitrogen, with the possible nitrosated product identified.

Methods

Knowledge-based scoring with 3 activating and 12 deactivating features · Graph Attention Network


Challenge

Nitrosamine impurities led to recalls and withdrawals of drugs including valsartan, ranitidine and varenicline, and FDA guidance for manufacturers is to assess nitrosamine risk in their products.


Approach

NitroSafeAI scores each nitrogen atom two ways: rules that encode known nitrosation chemistry, and a graph network trained on measured nitrosation yields.


Evidence

84% accuracy on a public benchmark of 437 nitrogen atoms in 207 molecules (Reiss et al., Chem. Res. Toxicol. 2025).

Outcome: A ready-to-deploy in silico tool that flags which nitrogen atoms in an API or excipient can form a nitrosamine

Drug product · Excipient selection

API surrogate search: Formulation starting points for a novel API

SitepKa output for one molecule in DMSO: A predicted pKa for each ionizable site. 2 of its 3 ionizable sites are shown.

Methods

Structure and property-based cheminformatics search · FDA formulation records · Data mining using Agentic AI


Challenge

A novel API has no formulation history, so the first excipients are often chosen by analogy and expertise.


Approach

The portal collates physicochemical data for ~3K approved APIs, such as melting point, pKa, logD and solubility, along with their inactive ingredients of ~150K FDA-approved formulations. Scientists can search by name, SMILES, structure or a property profile.


Evidence

Client engagement. The surrogate search model was deployed for a client’s formulation team in Europe.

Outcome: The approved drugs most similar to a novel API, and the excipients their formulations used, as starting points

See Igniva™ for Chemistry on your own problem

Bring the question you are working on, and we will show you which applications and agents apply.