Home / Directory / AI infrastructure / Inference & model hosting / TypeSafe AI
TypeSafe AI
San Francisco lab whose first model, Jev, returns structured decisions for other software, with a probability and a confidence score, and has been open to developers without a waitlist since September 2026. TypeSafe raised an $870 million Series A in October 2026, led by Andreessen Horowitz at a $7.5 billion valuation.
Developers who need a fast yes-or-no, a choice from a list, or a score inside a high-volume workflow, and who pay per input token.
Teams that need a chat model for drafting, or a model that takes audio or video.
Verdict
TypeSafe AI is a San Francisco lab founded in 2024 by Diogo Almeida, Erik Gafni, and Sasha Sheng. Almeida previously worked on reinforcement learning from human feedback, InstructGPT, ChatGPT, and GPT-4 at OpenAI. The first model, Jev, takes structured questions and returns answers other software can branch on, such as a yes-or-no probability, a selection from a list, or a score, each with a confidence measure. Developers set a threshold for when the application proceeds, asks for more information, or hands the case to a person. The company points Jev at high-volume steps such as classifying a service request, evaluating an invoice, triaging a security alert, or checking another agent's result.
SiliconANGLE reported on September 16, 2026 that TypeSafe emerged from stealth with a $40 million seed round led by DCVC. The company says Jev returns results in under 100 milliseconds and can be up to 100 times faster and less expensive than other frontier models. SiliconANGLE says the company's site listed a price of 39 cents per 1,000 workflows at launch. TypeSafe ended the waitlist in September 2026, and its documentation now lists $42 per billion input tokens, with output tokens free. DCVC general partner James Hardiman described the problem as making models reliable enough to embed in products at scale (SiliconANGLE). A same-day funding roundup also records the $40 million seed led by DCVC (TechStartups).
Score Breakdown
How TypeSafe AI scores in the categories that matter to its buyers.
Pricing
TypeSafe’s documentation lists Jev at $0.042 per million input tokens ($42 per billion), with output tokens free. At launch, SiliconANGLE reported a price of 39 cents per 1,000 workflows (SiliconANGLE).
The Field at a Glance
Where TypeSafe AI ranks among Inference & model hosting vendors we reviewed, by Overall Score and relative typical engagement cost.
TypeSafe AI scores 7.0 in this peer set, below Hyperstack (7.2), Together AI (7.3), and Modal (8.4). Relative cost sits at the low end of inference pricing at $42 per billion input tokens, and the waitlist closed in September 2026.
Use-case matrix
| Use case | Fit | Notes |
|---|---|---|
| Yes-or-no and scored decisions inside an application | Strong | Jev returns typed answers and a confidence measure. |
| Invoice checks, request classification, alert triage | Strong | Those are the high-volume steps the company lists. |
| Hundreds of decisions from one prompt | Mixed | The company says parallel outputs are part of the design. |
| Drafting email or chat for a person | Weak | The model is built for software, not a conversation. |
| A dependency you can call in production tonight | Weak | Jev has been open since September 2026, so teams have only weeks of production history to judge. |
Review Excerpts
Representative buyer views drawn from public information about TypeSafe AI. Paid reviews and pay-for-play sites were excluded.
“Jev returns a typed decision with a confidence score, so the code can proceed on high confidence and route the rest to a person.”
Representative buyer view
“We would point it at invoice checks and alert triage, where each decision feeds a rule in code.”
Representative buyer view
“The model is only weeks old, so we would test whether its confidence scores match accuracy on our own data before relying on it.”
Representative buyer view
Where customer reviews couldn't be found, our analysts attempted to give a voice to buyer sentiment, drawing on real information from review sites, forums, and product documentation.
Based on: TypeSafe’s model documentation and TechCrunch’s Series A report
Company Facts
- Founded
- 2024 (CB Insights)
- Headquarters
- San Francisco, California, United States (CB Insights)
- Key investors
- DCVC, Andreessen Horowitz, Sequoia Capital (CB Insights)
- Website
- typesafe.ai
Key Leadership
Diogo Almeida
Co-founder and CEO
Methodology
This page is an independent evaluation of TypeSafe AI for buyers comparing options in inference & model hosting. AI Industry Reviews accepts no sponsorships, advertising, or pay-for-placement fees. TypeSafe AI did not pay for this review.
What we scored
The headline number is an Overall Score on a 0-10 scale. Eight criteria fall under it in two groups.
- Buyer outcomes (for inference & model hosting)
- Structured decisions for software
- Latency claims
- Availability beyond a waitlist
- Fit inside repeated workflow steps
- Company & commercial
- Innovation & product leadership
- Project management & communication
- Pricing
- Contract fairness
Pricing measures whether the price looks fair for the value delivered, including packaging and renewal friction that show up in real buying cycles.
Score Composition
| Input | Weight | What it covers |
|---|---|---|
| Reviews | 40% | A proprietary read of what practitioners say about likes, complaints, and day-to-day use, including public review sites, forums, and private chat rooms. Paid reviews and pay-for-play sites are out of scope. |
| Product | 35% | Hands-on look at screens and workflows. |
| Pricing | 15% | Whether the price looks fair for what you get. |
| Docs & training | 10% | Docs, tutorials, and training material. |
How we balanced the evidence
The Overall Score is the simple average of the eight criteria. Recommendation language follows that score and the use-case matrix above.
Scope
TypeSafe AI is graded here as inference & model hosting. Criteria scores can move as more review volume and product checks are added.