AnGel AI: Software for Biomaterials R&D
Tell AnGel AI what you're building. It researches the field, designs formulations, plans the experiment and models your results, all in one chat.
-
Research
A sourced report across literature, patents, clinical, regulatory and market records.
-
Design
Ranked formulation concepts, each ingredient traced to a measured recipe and checked against FDA inactive-ingredient limits.
-
Execute
A plate-ready experiment and materials list, run in your lab or by GelTech on Carousel.
-
Model
Results build models from your own data, so each run plans a better next one.
An injectable hydrogel that releases a GLP-1 drug over 4 weeks, with degradation as the main release mechanism and minimal burst release.
Scout scanned 5,000+ sources across literature, patents, clinical trials, regulation and market. Report ready.
Know what's been done before you mix anything
Get a sourced research report across literature, patents, clinical trials, regulation, market and similar products through Scout, with every key finding marked verified or unverified.
- Literature18 records used
- Patents14 records used
- Clinical12 records used
- Regulatory and standards9 records used
- Market and funding7 records used
- Similar products11 records used
- A single-injection hydrogel depot sustained release for 42 days in rodentsVerified
- An issued patent covers a thermoresponsive injectable hydrogel depotVerified
- No clinical trial yet for this depot classUnverified
Start from formulations with evidence behind them
Get ranked formulation concepts through Design, each with its likely failure, the first test to run, and every ingredient traced to a measured recipe or flagged as an estimate.
- PolymerModified HPMC1 wt%Measured · recipe 4
- PolymerPEG-PLA nanoparticles10 wt%Measured · recipe 4
- DrugGLP-1 receptor agonist2.5 mg/mLMeasured · recipe 13
- ProcessSterile filtration of each streamEstimate
Plan a screening experiment, plate by plate
Lay out hundreds of formulations across well plates through the experiment planner, so one run maps the space instead of testing one recipe at a time.
- Built for Carousel. Plate formats, deck slots and the properties Carousel measures.
- Checked before it runs. The planner flags missing replicates and confirms inputs vary independently.
- Every property you need. Swelling, degradation, rheology, mechanics, structure and release.
Run it yourself, or have GelTech run it
Download the plan, the formulation CSV and the bill of materials to run the experiment in your lab, or send it to GelTech for an indicative quote and an ML-ready data package with QC.
- Experiment plan
- Formulation CSV
- Bill of materials
- Indicative quote
- Run on Carousel
- Data package with QC
Build models from your own data
Find out whether your data is ready to model through Data & Models: upload a dataset, map each column, and run a quality and predictability check.
Shown: a demo model on synthetic data, as in the app. Your own models are built once your data is ready to model.
Your data stays yours
- Encrypted in transit and at rest, and isolated per project
- Never used to train shared models
- IP, formulations and results stay client-owned under strict NDAs
Apply for the AnGel AI beta
AnGel AI is in beta for R&D teams who'll share feedback as it grows. Tell us what you're building and we'll set up a call and a walkthrough.