AI for biopharma — built by the peoplewho built the bench.
Thoth's four domains are the four things scientists still do every day.
Knowledge
Thoth was the god of writing and letters. Today: search the literature, read the evidence, turn scattered sources into an answer you can cite.
Reckoning
He was also the god of measure and counting. Today: analyze experimental data, run structure prediction and docking, design and model DoE.
Science
He was held to be the keeper of knowledge and craft. Today: design experiments, assess targets, support the whole path from hypothesis to evidence.
Record
He was scribe to the gods, keeper of everything written down. Today: process records, regulatory documents, an audit trail you can follow.
- 英 — Intelligence — and a nod to the parent company, 英赛斯 (Inscinstech).
- 索 — Exploration — to search, to verify, to find the route.
- 思 — Thought — to reason, not merely to restate.
"In-" is the prefix shared across the Inscinstech product family; "Thoth" is the god's name. Together they say the whole thing: an agent that does knowledge, reckoning, science and record on a scientist's behalf.
We did not start with an AI and go looking for a use for it.
Inscinstech spent nine years building biopharma instruments. Our engineers live on the process floor, and they kept seeing the same problems — none of which a general-purpose chatbot solves.
Generic AI does not know CMC
It has read the papers but never run a purification. It cannot tell you what this resin costs you in yield at this pH, or which route falls apart at pilot scale.
Sensitive process data cannot leave the region
Yet most AI tools send it overseas by default. For teams working toward preclinical and filing, that one fact rules out an entire class of tooling.
Scale-up runs on trial and error
The industry still crosses the bench-to-pilot cliff by spending batches and months. What actually predicts it is continuous, same-source data across scales — and you need instruments to have that.
Those three only get solved together, and only by people holding the instruments, the process data and the AI at the same time. That is what InThoth was built for.
Why this, why now.
- LLM economics have crossed the threshold for biopharma-specific deployments.
- Domestic compute and compliant hosting have matured — sensitive process data can stay in-region.
- Generic AI tools don't understand CMC. We do.
- The biopharma industry needs faster, more compliant tooling — and the current incumbents are not innovating.
How we work.
Domain truth
We build for biopharma, not generic enterprise SaaS.
Compliance from day one
Audit-ready by default, not retrofitted.
Customer data is not training data
Period. Contractually guaranteed.
Bilingual by default
Global product, CN data residency.
InThoth is built by Inscinstech.
Suzhou Inscinstech has been building biopharma instruments and software since 2017 — bioreactors, nucleic acid synthesis, protein chromatography, tangential flow filtration, analytical instruments, every line spanning laboratory through production scale. Here is why that matters: what we train and calibrate InThoth on is the real process data those instruments produce, plus several hundred bioprocess development cases. A software-only team cannot get at that.
In the news.
- 2026-05-12The InThoth platform launches.
- 2025-12-30By the end of 2025, raised several hundred million RMB in total.
We're hiring.
Engineers, process scientists, bio / chem PhDs, sales, product — almost every role is open.
See open positions