Question 1

What is Sapiens Technology?

Accepted Answer

Sapiens Technology is a Brazilian artificial intelligence company founded by Ben-Hur Varriano and Belchior Varriano. It is the first Latin American company with a proprietary multimodal LLM, and the creator of the SAPI (Schizophrenic AI) models. It offers three products: Sapiens Builder, a platform for companies to train their own foundational models; Sapiens Chat, for advanced conversations with AI; and the Sapiens API, for developers, currently in final development.

Question 2

What is Sapiens Builder?

Accepted Answer

Sapiens Builder is Sapiens Technology’s flagship product: the platform that lets companies and individuals build their own foundational artificial intelligence models, trained on their own data. It uses the proprietary SAPI architecture, far more efficient than conventional training techniques, which makes CPU-level training viable. Training can run on the organization’s own hardware or on Sapiens infrastructure, and both the data and the resulting model stay under the customer’s control.

Question 3

What is Sapiens Chat?

Accepted Answer

Sapiens Chat is the first Latin American AI chatbot with a proprietary multimodal LLM, developed in Brazil. What sets it apart is perpetual context memory — conversations are unlimited with no token ceiling, and the system recalls every earlier interaction. It includes the SAPI models from SAPI-1 to SAPI-5.1, GitHub repository import, and support for multiple formats. It is available on the Free, Starter, Local, Business and Enterprise plans.

Question 4

How does Sapiens Chat compare to ChatGPT?

Accepted Answer

Where ChatGPT (OpenAI) uses a conventional GPT Transformer architecture, Sapiens Chat uses the proprietary SAPI (Schizophrenic AI) architecture, which evaluates multiple internal responses before returning the best one to the user — giving a higher degree of refinement. The main difference is perpetual context memory: Sapiens recalls every earlier conversation with no token limit, while ChatGPT has bounded context windows. Sapiens Chat is developed 100% in Brazil, on its own model, with no dependency on third-party APIs.

Question 5

How does Sapiens Chat compare to Claude (Anthropic)?

Accepted Answer

Sapiens Chat and Claude (from Anthropic) are both advanced AI chatbots, but with different architectures. Claude uses a conventional Transformer with a focus on safety and alignment. Sapiens Chat uses the SAPI (Schizophrenic AI) architecture with internal multi-model evaluation. What sets Sapiens apart is being developed in Brazil for the Latin American context, with perpetual context memory, and with fine-tuning available through the Sapiens API. Sapiens is the Brazilian alternative with a proprietary LLM.

Question 6

What is the Sapiens API?

Accepted Answer

The Sapiens API is Sapiens Technology’s developer platform. It allows two paths: (1) fine-tuning existing SAPI models on your own training data; (2) building an LLM from scratch on the proprietary SAPI architecture. The API matches the industry REST standard and provides management of organizations, projects and API keys in the sk-* format. The SAPI architecture is far more efficient than conventional training techniques. The Sapiens API is in final development: access is not yet open to the public and sign-ups run through a waitlist.

Question 7

What is the SAPI (Schizophrenic AI) architecture?

Accepted Answer

SAPI (Schizophrenic AI) is the proprietary AI architecture developed by Sapiens Technology. Inspired by the mental processes of schizophrenic geniuses, the base model generates a primary response and, before returning it, receives outputs from auxiliary models for internal evaluation — picking the most refined answer. The Frankenstein manager orchestrates sub-models specialized in text, code, image, audio and video, which makes the system genuinely multimodal. The Master Entity integrates everything into a single unified matrix. This approach produces higher-quality responses at a lower computational cost than conventional GPT transformers.

Question 8

Is Sapiens available in Brazil?

Accepted Answer

Yes. Sapiens Builder, Sapiens Chat and the Sapiens API are products of Sapiens Technology LTDA, a Brazilian company. All of them are developed in Brazil and serve the Brazilian and Latin American markets first. Sapiens Chat has plans priced in Brazilian reais (BRL) and support in Portuguese. The platform is available in Portuguese, English and Spanish.

Question 9

How do you fine-tune an LLM with the Sapiens API?

Accepted Answer

To fine-tune with the Sapiens API: (1) create an account and get your API key (sk-*); (2) upload your training data through the /v1/fine-tuning/upload endpoint; (3) create a fine-tuning job, choosing the base model (SAPI-1 to SAPI-5); (4) once training finishes, the customized model is available for inference through the REST API. The fine-tuned model inherits the SAPI architecture — including internal multi-model evaluation. You can reach it with the identifier ft:sapi-4:org::ckpt-id.

Question 10

Which models are available on Sapiens?

Accepted Answer

The SAPI model family spans six generations: SAPI-1 (built for speed, fast responses at low computational cost), SAPI-2 (broader cognitive capacity for everyday tasks), SAPI-3 (deep reasoning for data analysis and development), SAPI-4 (the full capacity of the base model, the frontier of Sapiens research), SAPI-5 (state-of-the-art extended reasoning) and SAPI-5.1 (the generation in production today, with the same extended reasoning as SAPI-5 at a far lower computational cost). In Sapiens Chat, the Free plan includes Sapi-5.1 micro; the paid plans, from Starter up, add Sapi-5.1 tiny-turbo.

Question 11

Does Sapiens Chat remember earlier conversations?

Accepted Answer

Yes. Sapiens Chat has perpetual context memory — one of the platform’s most distinctive features. The system recalls every earlier conversation, however many dialogues there are and however long the interactions run, with no token limit. This is possible thanks to the morphological storage algorithm of the Frankenstein class, which saves the binaries of each interaction into indexed files that are consulted instantly on every new prompt.

Choose your plan

Sapiens API pricing