Who uses TypeSafe AI?
TypeSafe AI builds System One models.
usedby has found 2 public customers of TypeSafe AI so far: Jack & Jill and MotherDuck. 1 of them is confirmed by a published case study, quote or press release. 1 of them is shown by TypeSafe AI on its own pages. Every company links to the public page that names it.
2 companies, each with its public proof
Strongest proof first. Open a company to see the page that names it, and when we captured it.
ConfirmedJack & JillSource: typesafe.ai · Captured Oct 27, 2026
ObservedMotherDuckData & AnalyticsSource: typesafe.ai · Captured Oct 27, 2026
| Company | Industry | Proof | First public proof | Source |
|---|---|---|---|---|
| Jack & Jill | — | — | typesafe.ai | |
| MotherDuck | Data & Analytics | — | typesafe.ai |
What customers say TypeSafe AI changed
Figures and quotes as TypeSafe AI and its customers published them, each tied to its source. We quote, we do not measure the result.
“We saved ~half a million dollars last week with Jev.”
Matthew WilsonCo-founder/CEO, Jack & JillWhat TypeSafe AI does, and what to check
What TypeSafe AI does
TypeSafe AI builds System One models. Its model, Jev, does not generate text: it reads a state and answers typed questions (a choice, a score, a yes/no, a box on an image) with calibrated probabilities and a confidence level, in 70 to 500 ms. It is made for high-volume decisions inside software: classification, routing, candidate screening, guardrails and re-ranking. You pay per input token ($0.042 per million) and output tokens are free. Jev launched in early access in September 2026 and is also served through Vercel AI Gateway, Netlify AI Gateway and OpenRouter.
- Typed answers: Choice, Score, yes/no and image Box
- Calibrated probabilities and a confidence level per answer
- Many questions answered in parallel against one state
- 70 to 500 ms end-to-end latency
- Image inputs (Jev 1.14)
- Python and JavaScript SDKs
- Candidate screening
- Classification and intent routing
- LLM guardrails
- RAG passage filtering and re-ranking
- Structured data extraction
- Citation checking
- Vercel AI Gateway
- Netlify AI Gateway
- OpenRouter
- LangChain
- Pydantic AI
- Langfuse
- MotherDuck
Written by usedby from typesafe.ai.
Security
- SOC 2 Type II
Pricing and terms
- Pricing
- Jev 1.14: $0.042 per Mtok
- Your data
- Not used to train models by default.
Read on trust.typesafe.ai, checked Oct 27, 2026
Read on docs.typesafe.ai, checked Oct 27, 2026
Read on typesafe.ai, checked Oct 27, 2026
Questions to ask in your demo
- Which of your customers look like us, by industry and size?
- What does a pilot cost, and what decides the move to a full contract?
- Where is our data stored, who can access it, and is it used to train models?
- How does TypeSafe AI work with the tools we already use?
TypeSafe AI alternatives, by public customers
Same or competing category, from another vendor. Every proof level counts. A shared category alone does not establish the same functions or suitability for your task.
- TypeSafe AIThis tool
ClaudeAI assistant for advanced reasoning, coding, and content creationWho uses Claude →
GeminiGoogle's family of multimodal AI models and the Gemini assistant app, formerly BardWho uses Gemini →
BasetenInference platform to deploy and autoscale open-source, custom and fine-tuned AI models.Who uses Baseten →
ChatGPTVersatile AI chatbot for conversation and tasksWho uses ChatGPT →
Nebius AI CloudAI cloud with non-virtualized NVIDIA GPU clusters, MLOps tooling and Token Factory managed inference.Who uses Nebius AI Cloud →
Questions about TypeSafe AI
Who uses TypeSafe AI?
2 companies use TypeSafe AI (1 confirmed, 1 observed), including Jack & Jill and MotherDuck.
What are the alternatives to TypeSafe AI?
Ranked by public customers (every proof level counts, job posts included), the top TypeSafe AI alternatives are Claude (179 customers, 0 shared), Gemini (79 customers, 0 shared), Baseten (66 customers, 0 shared), and ChatGPT (55 customers, 0 shared). See all TypeSafe AI alternatives.
How does usedby know who uses TypeSafe AI?
Each company is linked to a public page that names it: TypeSafe AI's customer page, a published case study, a press release or a job post. We save most pages with a dated capture and check them again on each run. A company leaves the list only after three checks in a row that no longer find it, and at least 14 days after it was last seen. A company with a published case study or quote is never removed automatically.
What do confirmed, observed and signal mean?
Confirmed (1): a case study, a named quote or a press release. Observed (1): TypeSafe AI shows the company on its own pages, for example a logo wall. Signal (0): a job post shows the company’s employees use TypeSafe AI, the weakest proof. Every level counts as a customer, and each line shows its level.
How current is this list?
The most recent check was on October 27, 2026, and 2 of 2 companies on this list have been checked so far. A company leaves the list only after three checks in a row that no longer find it, and at least 14 days after it was last seen. A company with a published case study or quote is never removed automatically.
Where this data comes from
Two proof levels
Cite this page
usedby (2026). Who uses TypeSafe AI? 2 companies. Retrieved from https://www.usedby.ai/tools/typesafe-ai, latest check October 27, 2026.