IntelliForm is an open-source agentic AI framework designed to accelerate specialty chemical R&D. It translates natural language requirements into high-precision, RDKit-validated chemical formulations and manufacturing protocols.
- Full Platform (Streamlit Lab): The core R&D cockpit for internal research, lab-scale prototyping, and deep optimization.
- IntelliForm Free (Web): Public production app featuring Google Sign-in and rate-limited QSAR generation.
- Interactive Demo: A guided showcase of agentic reasoning in chemical manufacturing.
- Official Project Site: Documentation and ecosystem overview by ChemeNova.
- Natural Language → Chemistry: Convert briefs into structured SMILES and formulation tables using LLM-based parsing.
-
Multi-Objective Optimization: Balance assay purity (e.g., target 98–99%
$CaCl_2·2H_2O$ ), cost, and sustainability metrics. - Regulatory Intelligence: Integrated QSAR modeling and regulatory screening (v0.9+).
- Agentic Reliability: Local JSON normalization and contradiction filtering to ensure "pilot-ready" outputs.
- Manufacturing Digital Twin (v1.6): Simulate the process timeline, energy/CO2, yield, and scale-up risks for a formulation from lab → bench → pilot → production batch sizes.
- Self-Driving Lab (SDL) Integration (v1.6): Generate liquid-handling protocols (Opentrons OT-2 / generic) for automated lab hardware, then ingest measured results to auto-detect spec failures and propose the next blend via Closed-Loop Reformulation Intelligence™.
This repository supports a dual-stack architecture to separate research from public access:
| Feature | Streamlit Lab (app.py) |
Public Web/API (web/ + api/) |
|---|---|---|
| Primary Use | Internal R&D / Heavy Computation | Public Showcase / User Onboarding |
| Tech Stack | Python + Streamlit + RDKit | React (Vite) + FastAPI + Supabase |
| UX Focus | Complexity & Scientific Tooling | Performance, Auth, & Responsiveness |
| Auth | Local/Session | Google Sign-In |
conda create -n intelliform python=3.11 -y
conda activate intelliform
conda install -c conda-forge rdkit -y
pip install -r requirements.txt
streamlit run app.pypip install -r requirements.txt
uvicorn api.main:app --reloadnpm install
npm run devMIT — Distributed by ChemeNova in collaboration with ChemRich Global. Built for chemists, ML engineers, and sustainability builders. 🚀