CheckTOS

Jev

by TypeSafe AI, Inc.


TypeSafe, maker of Jev

TypeSafe AI keeps two sets of terms. The Master Customer Agreement governs the product. The Terms of Use governs the website. Both are public, and you can switch between the two reads below.

Clarity Meter Needs a lawyer

This is a full commercial agreement, dense with defined terms and long all-caps blocks for liability, disclaimers, and arbitration.

Plain EnglishSome legaleseDense legaleseNeeds a lawyer
Updated: 19 Sep 2026
Retrieved: 23 Sep 2026
Also read: Privacy Policy, Data Processing Addendum

The agreement that actually governs Jev keeps your inputs and outputs yours, and TypeSafe will not train on your data without your consent.

  1. TypeSafe does not claim your Input and assigns you the Output, so what you put in and get out stays yours.Your work
  2. TypeSafe will not put your Customer Data into a training set without your prior consent, and the Privacy Policy makes the same promise for your Input.Your data
  3. Liability on both sides is capped at the greater of twelve months of fees or fifty dollars, and indemnity runs both ways.Your legal rights
  4. Disputes go to individual arbitration with a class-action waiver and no opt-out, under California law.Your legal rights

Product

The Services are the TypeSafe-hosted console at console.typesafe.ai and the API. You get a limited license to use them for the length of your deal. On ownership the agreement is plain: TypeSafe does not claim your Input, disclaims the Output, and assigns the Output to you. TypeSafe keeps everything on its own side, the Services, the Telemetry, and the technology behind them.

"TypeSafe grants to Customer a limited, non-exclusive, non-transferable, non-sublicensable license during the Term to: (a) access and use the Services in accordance with the applicable documentation made available by TypeSafe from time to time"Section 2.1 · License
"TypeSafe does not claim ownership of Input and TypeSafe disclaims ownership of Output. TypeSafe hereby assigns to Customer all of its right, title, and interest, if any, in the Output."Section 4.2 · Output

TypeSafe and its licensors keep all rights in the Services, Documentation, Telemetry, and TypeSafe technology, and any product feedback or suggestions you send, it "may use and exploit... without restriction or obligation" (Section 11).

Usage

Day-to-day use comes with hard limits. You cannot use the Services or Output to distill or train a competing model, reverse engineer them, or exceed the Usage Limits. You carry the Input: its content, its accuracy, and every consent needed for TypeSafe to process it. Where your Input holds personal data, the Data Processing Addendum makes TypeSafe a processor that acts only on your written instructions.

"Customer will not do (and will not attempt to do)... (b) use the Services or any Output... to perform model distillation, train a model to imitate the output of the Services, or develop (or to facilitate the development of) a similar or competing product or service"Section 2.3 · License Restrictions
"Customer is responsible for Input, including its content and accuracy, and will comply with Laws when using the Services."Section 5 · Customer Obligations

On personal data the Data Processing Addendum makes TypeSafe a processor that "will only Process Customer Personal Data to provide the Services and in accordance with Customer's documented instructions," will not sell or share it, gives you fifteen days to object to a new subprocessor, and reports a security incident within seventy-two hours (Data Processing Addendum, Sections 2.1, 2.2, 3.2, 5.2). TypeSafe may suspend access for a breach, a payment thirty days overdue, a change in law, or risk to other customers (Section 6).

Feedback & improvement

This is where the product terms stand apart. TypeSafe will not include your Customer Data in a training set without your prior consent, and the Privacy Policy repeats the promise for your Input. The one wide opening is Telemetry: the logs, hashes, statistics, and "learnings" about how you use the Services, which TypeSafe may process without restriction and keep for good.

"The foregoing license does not grant TypeSafe the right to, and TypeSafe will not, include Customer Data in a dataset used to train (i.e., to modify the model weights of) any artificial intelligence or machine learning models without Customer's prior consent."Section 4.1 · Use of Customer Data
"We will not train or fine tune any artificial intelligence or machine learning models on your prompts or other Input."Privacy Policy · Personal Data We Collect

The Privacy Policy also says TypeSafe "will not disclose any Input to a third party other than our service providers," and that it may make "anonymized or aggregated data that we may use for lawful purposes" (Privacy Policy). The training bar turns on "Customer's prior consent," and the agreement does not say how that consent gets asked for or given (Section 4.1).

Structure summary

Each item below is quoted and located in the documents.

One-sided
2 found
Neither party may assign, "except that TypeSafe may assign this Agreement in connection with a merger, reorganization, acquisition, or other transfer of all or substantially all its assets or voting securities" (Section 16.1). TypeSafe may suspend access on its own initiative, with prior notice only where practicable (Section 6).
Absolute
2 found
An "AS IS" and "AS AVAILABLE" disclaimer (Section 9.3). On termination TypeSafe "may delete Customer Data at any time in its sole discretion" (Section 10.3).
Undefined terms
2 found
"materially" sets the service warranty and "misuse" carves it out (Section 9.1); neither is defined.
Scope limits
3 found
Arbitration excludes small-claims, agency, injunctive-in-aid, and IP suits and waives class actions with no opt-out (Sections 15.2, 15.7); the Data Processing Addendum caps international transfers and gives a fifteen-day subprocessor objection (Sections 3.2, 6); the service warranty reaches only conformance with the Documentation (Section 9.1).
Gaps
3 found
The Order that carries the Term, Fees, and Usage Limits is not public; the consent mechanism for training on Customer Data is not described (Section 4.1); no fixed retention period, since Customer Data may be deleted at TypeSafe's discretion (Section 10.3).
Contradictions
0 found
None found.

How to reach them

Taken from the documents and sorted by what you need.

Support
support@typesafe.aiTerms of Use · Section 13(e)
Privacy and data
privacy@typesafe.aiPrivacy Policy · Contact Information
Legal notice
TypeSafe AI, Inc., 255 California St, Suite 1300, San Francisco, CA 94117. Sent by certified U.S. Mail or Federal Express with signature required.Master Customer Agreement · Section 15.4
Arbitration opt-out
By letter to TypeSafe AI, Inc., Attention: Legal Department – Arbitration Opt-Out, 255 California St, Suite 1300, San Francisco, CA 94117, within thirty (30) days after you agree to the Terms. This opt-out belongs to the website Terms of Use. The product Master Customer Agreement waives class actions and gives no opt-out.Terms of Use · Section 12(c)

Caveats

This explains what the documents say. It does not read the law for you, weigh whether a clause would hold up, or guess what happens if you break the terms. It is not legal advice, and it does not judge TypeSafe. Every quote is word-for-word from its source: the Master Customer Agreement and Terms of Use updated 19 September 2026, the Privacy Policy updated 19 November 2025, and the Data Processing Addendum updated 24 April 2026. The Terms of Use was read on 22 September 2026, and the Master Customer Agreement, Privacy Policy, and Data Processing Addendum on 23 September 2026. The separate product agreement is the public Master Customer Agreement, now read and folded in, so this covers both the product and the website terms.

Read another breakdown

Read the fine print on anything you sign.

CheckTOS is a downloadable skill. Point it at any terms of service and get this same plain-language read inside your own Claude, on your own tokens.

Download Skill ↓