QSPR Toolbox QSPR & QSAR Modelling Software for Research Labs in India
QSPR Toolbox is integrated QSPR software that gives you one managed workspace for Quantitative Structure–Property Relationship (QSPR/QSAR) work. As a result, you can go from molecular structures to descriptors, graph-theoretic topological indices, validated machine-learning models and property predictions in one place. Instead of assembling and maintaining a dozen separate tools yourself, you get a single setup that is built, deployed and supported by Siri Techno Fabs, Bengaluru.
Predict molecular properties from structure, reproducibly
QSPR models relate a molecule's structure, encoded as numerical descriptors and topological indices, to a measurable property such as boiling point, solubility, logP or toxicity. Similarly, QSAR (Quantitative Structure–Activity Relationship) applies the same approach to biological activity. The open-source ecosystem for this work is excellent. However, it is scattered across Python, Java, C++ and Perl, and each tool has its own installation, input format and version conflicts. QSPR Toolbox therefore brings these engines together as a single QSPR software installation and workflow. In addition, our team sets it up, configures it and supports it for your lab.
One install, many engines
For example, RDKit, Mordred, PaDEL, CDK, Open Babel, Chemprop and DeepChem are version-pinned and tested to work together in one environment.
Descriptors to models in one pipeline
First, the pipeline cleans your structures and calculates descriptors and indices. Then it handles feature selection, model training, validation and applicability-domain checks, and it logs every step.
Built for research and reporting
Because descriptor matrices, model metrics and run logs are all exportable, they directly support thesis chapters, journal methods sections and OECD-principle-aligned model documentation.
A complete QSPR workflow
Every project follows the same six steps. First, you import your structures; finally, you validate the model and export a report.
Integrated engines
QSPR Toolbox builds on proven, peer-reviewed open-source projects. In addition, each one is credited to its original developers, with a link to its official source.
Molecular descriptor & topological index calculators 5 engines
Mordred / Mordred-Community
It calculates around 1,800 molecular descriptors, for example Zagreb indices, Wiener index, Balaban J, detour matrix, adjacency/distance-matrix descriptors and autocorrelation descriptors. Because the community fork is actively maintained, the toolbox uses it.
PaDEL-Descriptor & PaDELPy
It provides 1,875 descriptors (1,444 1D/2D and 431 3D) plus 12 fingerprint types, built on the Chemistry Development Kit. It is also the descriptor engine behind OPERA.
RDKit
It handles molecular graph operations in 2D and 3D, Morgan/ECFP, MACCS and topological-torsion fingerprints, and 200+ physicochemical descriptors. Therefore, it forms the structural backbone of the toolbox.
CDK (Chemistry Development Kit)
It covers structural cheminformatics and graph isomorphism. In addition, it computes Wiener, Zagreb and Kier–Hall topological indices and generates QSPR descriptors.
MayaChemTools
It generates topological pharmacophore fingerprints, atomic invariants and physicochemical descriptors. Moreover, its scripts are well suited to batch processing of large compound sets.
+ STF Custom Index Module
We implement and validate degree-based and distance-based topological indices that the standard packages do not include. For the full list, see the section below.
Chemical structure handling 1 engine
Open Babel & Pybel
It converts between 110+ chemical file formats (SMILES, SDF, InChI, MOL2 and more). It also generates 3D conformers and then runs force-field energy minimisation before descriptor calculation.
QSPR / QSAR modelling frameworks 3 engines
QSPRpred
It builds QSPR/QSAR and proteochemometric (PCM) models. Furthermore, it integrates RDKit, Mordred, PaDEL, Mold2, scikit-learn, PyTorch and Chemprop under one reproducible model-serialisation format.
QSPRmodeler
It combines RDKit and Mordred descriptors with automated machine-learning cross-validation pipelines. As a result, it suits quick molecular predictive analytics.
ZairaChem
It is an automated machine-learning pipeline for QSAR/QSPR property prediction from chemical structures. In other words, it works out of the box with minimal configuration.
Deep learning & graph neural networks 3 engines
Chemprop
Developed at MIT, it predicts molecular properties directly from the molecular graph with D-MPNN models. In addition, QSPRpred supports it natively.
DeepChem
It offers graph-convolution and deep-learning models for molecular property prediction. Likewise, it covers quantum chemistry and materials science.
OpenChem
It provides graph convolutional networks and other deep-learning architectures. Consequently, it is useful for QSAR/QSPR and wider computational chemistry work.
Ready-made models, regulatory tools & model repositories 3 engines
OPERA (NIEHS)
It gives open-source predictions of physicochemical properties, environmental fate, ADME and toxicity endpoints. Because it uses PaDEL descriptors, it fits directly into the toolbox, and it also includes built-in applicability-domain assessment.
OECD QSAR Toolbox (interoperability)
Regulators and industry use it for chemical hazard assessment, property profiling, chemical grouping, read-across and QMRF reporting. QSPR Toolbox prepares and exchanges data with it. However, the OECD Toolbox itself is distributed by the OECD and installed separately, free of charge.
QsarDB
It is an open framework and repository of published QSPR/QSAR models, with Java/Python libraries and ONNX support. Therefore, you can reuse existing models and benchmark your own against them.
Web interface & REST API 1 engine
Topological indices, standard and custom
Many QSPR studies in mathematical chemistry rely on graph-theoretic indices, for example in drug-property ranking and polymer or nanostructure analysis. However, standard packages leave out many of these indices. QSPR Toolbox therefore computes the common indices out of the box. In addition, our R&D team can add newer or study-specific indices and then check each one against hand-computed reference graphs.
Hydrogen-suppressed molecular graphs
Indices are computed on clearly defined graph representations. Moreover, the convention used for each one is documented, so your values can be reproduced and compared with published tables.
Index–property regression
The toolbox fits linear, quadratic and multiple regression of indices against experimental properties. In addition, it reports R², RMSE, F, p-values and cross-validated Q² for every model.
Built for research labs and R&D teams
Universities & research centres
It serves mathematics, chemistry and pharmacy departments running QSPR, topological-index or drug-design projects. It also suits funded research projects that need a dependable computing setup.
PhD scholars & supervisors
You get a validated environment that produces descriptor matrices and model outputs. Therefore, you can describe and defend your results in a thesis or journal methods section.
Pharma, chemical & materials R&D
Use it for property screening, ADME/toxicity estimation and early-stage prioritisation on a private server. Meanwhile, your structures stay in-house.
Choose how you want it delivered
Every QSPR software option below includes installation and testing. In addition, you can combine any of them with training and support.
| Option | What you get | Best for | Starting from* |
|---|---|---|---|
| Workstation Setup | We install and test the toolbox on your Windows/Linux workstation. Then we walk you through a sample end-to-end project in a hands-on orientation session. | Individual researchers and PhD scholars | ₹3.5 Lakh |
| Lab Server | We set up a multi-user installation on your server, or on one we supply through our IT division. It also includes the web GUI/API, user accounts, GPU support for deep-learning models and scheduled backups. | Departments, research groups and funded projects | ₹6.5 Lakh |
| Private Cloud | We deliver a containerised deployment on your cloud account or a managed instance. In addition, it comes with access control and remote support. | Distributed teams and industry R&D | ₹8 Lakh |
| Custom Development | We build new topological indices, custom descriptors, dataset curation pipelines or bespoke models. We can also integrate the toolbox with your existing systems. | Projects with specific methodological needs | ₹4 Lakh |
| Training & Support | We run workshops for faculty and students and provide annual maintenance with version updates. Furthermore, you get priority troubleshooting. | Any of the above | ₹3.25 Lakh |
* Indicative starting prices, excluding GST. However, final pricing depends on deployment type, number of users, hardware and customisation. Therefore, request a quote below and we will send a scoped proposal.
Common questions
What is QSPR software used for?
QSPR software calculates numerical descriptors and topological indices from molecular structures. It then builds statistical or machine-learning models that predict physicochemical properties such as boiling point, solubility, logP or toxicity. As a result, it is widely used in drug design, materials research, environmental chemistry and mathematical chemistry.
What is the difference between QSPR and QSAR?
QSPR models predict physical and chemical properties from structure. In contrast, QSAR models predict biological activity or toxicity. However, the workflow (descriptors, model training, validation) is the same, so QSPR Toolbox supports both.
How much does QSPR Toolbox cost?
A single-workstation setup starts from ₹3.5 lakh (excluding GST). In addition, lab server and private cloud deployments, custom topological indices and training are quoted separately, based on users, hardware and scope.
Is QSPR Toolbox the same as the OECD QSAR Toolbox?
No. The OECD QSAR Toolbox is a separate regulatory application published by the OECD. By contrast, QSPR Toolbox is a Siri Techno Fabs research-software offering that integrates open-source QSPR engines. It can also exchange data with the OECD Toolbox, which you install separately.
Do you own the engines listed on this page?
No. The engines are open-source or freely distributed projects created by their respective developers and institutions, and we credit and link each one. Instead, our work is integration, configuration, validation, custom development, deployment and support. Furthermore, we deliver it under each component's own licence terms.
Will my molecular data leave my institution?
Not with the Workstation or Lab Server options, because everything runs on your own hardware. Similarly, for Private Cloud deployments, the data stays in the cloud account and region you choose.
Can you add a topological index from a recent paper?
Yes. Simply send us the definition or the paper. First, we implement the index; then we check it against reference graphs and document the graph convention it uses.
What operating systems and hardware are supported?
Linux is recommended for servers, and Windows workstations are also supported. Deep-learning engines such as Chemprop, DeepChem and OpenChem run much faster on an NVIDIA GPU. However, a GPU is not required for descriptor calculation or classical models.
Do you provide training?
Yes. We run orientation sessions with every setup. In addition, workshops for departments and research groups are available on request.
Related R&D solutions
Research & Development software
It covers all seven R&D areas, from battery management to software defined radio.
AI integration services
We also deliver private on-premise AI assistants, document chatbots, computer vision and OCR solutions.
IT solution & server support
Similarly, we supply servers, storage and networks to host your research computing environment.
Tell us about your QSPR project
Share a few details about your project. Then our R&D team will reply with a scoped proposal covering deployment, engines, customisation, training and timeline.
- Phone / WhatsApp: +91 81056 44422
- Email: support@siritech.net
- Office: 643, 68th Cross Rd, Kumaraswamy Layout 1st Stage, Bengaluru, Karnataka
Attribution and licences
RDKit, Mordred, PaDEL-Descriptor, PaDELPy, CDK, Open Babel, MayaChemTools, QSPRpred, QSPRmodeler, ZairaChem, Chemprop, DeepChem, OpenChem, OPERA, QsarDB, GenUI and the OECD QSAR Toolbox are the work of their respective authors and institutions. Furthermore, they are distributed under their own licences. Names are used only to identify the software. However, Siri Techno Fabs is not affiliated with or endorsed by these projects or the OECD. Also note that descriptor counts are as reported by each project and may change between versions. Contact us · All R&D software