AI Use Disclosure

Some parts of Techoclaro use an artificial intelligence (AI) system — a large language model. When a screen shows text written by AI, it says so, and it names the model and when the text was generated. When a screen uses a scripted answer instead, it says that no AI model was used. The assistant called “Techy” is an AI system, not a person.

AI output in Techoclaro is never legal advice, it does not decide anything for you, and it is never a reason to miss a deadline. Techoclaro LLC is not a law firm; see our Legal Information Disclaimer.

• Reading your documents. When you upload a photo, scan or PDF of a letter, an AI model transcribes the words on it, exactly as written. It does not summarize or interpret. The dates and deadlines are then found and calculated by our own rules-based software, not by AI.
• Techy, the case assistant. Answers questions about your case, using only the information stored in your case (such as the kinds of documents, dates, deadlines and findings) and the laws it can cite. It does not see other people’s cases.
• Explanations. Plain-language explanations of a step in your case map, of a right that may apply to you, or of a court decision in our library.
• Hearing practice. Coaching on a statement you practice for a hearing, based on the facts you typed.
• Screen guide. General answers about how the app works or how the process works in your state. It never answers questions about a specific case.
• Credit report reading (Credit Health). Helps organize the accounts in a credit report you upload. Before any text is sent to the AI, we remove Social Security numbers and cut account numbers down to the last four digits.

What does not use AI: calculating legal deadlines, the timeline, reminders, eligibility checks, prices and billing. Those follow fixed rules written from the law and our terms.

• Grounded in sources. The AI is given specific sources — your case information, laws, or court decisions from our library — and told to use only those. A statement that points to a law or court decision is checked against our records; a citation we cannot match is removed, not shown.
• No promises. The AI is instructed never to predict an outcome or promise a result, such as that a sale will stop or a loan will be modified.
• Human review. Most AI output is shown to you right away, without a person reviewing it first. Each AI panel says whether a person has reviewed it (and when).
• Records. For each AI run we keep a record of which feature ran, which model was used, when, and a fingerprint (hash) of the input and output, so we can investigate problems. We do not keep the AI’s full text in that record.

AI can make mistakes, and it can sound confident when it is wrong. It can misread a blurry page, misunderstand a question, or describe a law too broadly.

• Compare any transcription with the original document, especially names, dates and amounts.
• Verify every citation before you rely on it or put it in a letter or court filing: open the law or case and read it yourself. Laws change, and a law that applies in one state may not apply in yours.
• Check deadlines against the papers you received, and ask a lawyer, legal aid office or HUD-approved housing counselor when in doubt.

If you find an error, tell us at privacy@techoclaro.com. It helps us fix it for everyone.

Our AI features use models provided by Anthropic, PBC. When you use one of these features, we send the provider only what that feature needs:

• the document image or PDF you asked us to read;
• your question, the recent messages in that conversation, and the information stored in the one case you are asking about;
• the laws, court decisions or guide text the answer must be based on;
• for hearing practice, the facts you typed; for credit reports, the report text after Social Security numbers and full account numbers were removed.

We do not send your password, payment card details, or information from other people’s cases. The provider processes this data to return a response to us. [CONFIRM before publishing: the provider’s commitments under Techoclaro’s agreement — that inputs and outputs are not used to train its models, and how long it retains them.]

We do not use AI to make decisions that have legal or similarly significant effects on you, such as approving or denying anything. You can use the timeline, deadlines and reminders without asking the AI anything. Our Privacy Policy lists all of our service providers and your privacy rights.