# Biolevate

> Biolevate is an AI Knowledge Platform accelerating life sciences. Its core product, the Elise platform, connects to an organization's document storage backends, indexes documents with AI, and runs asynchronous Question Answering, Extraction, Multi-Dimensional (entity/tabular) Extraction, Find-Similar, and Agent jobs over them. Use cases target regulatory affairs, R&D, industry, and post-market workflows: systematic literature review, Clinical Study Report (CSR) and CTD assisted generation, batch-record quality control, results extraction, HTA support, pharmacovigilance PSUR updates, and regulatory intelligence. This llms.txt is generated by the API Evangelist enrichment pipeline from the public Biolevate developer surface; the provider does not publish an official llms.txt.

## API
- [Biolevate API (OpenAPI 3.1)](https://raw.githubusercontent.com/Biolevate/biolevate-api-sdk/main/openapi/biolevate-api.json): Core REST API for the Elise platform — 52 operations across Providers, Provider Items, Files, Collections, Question Answering, Extraction, Multi-Dimensional Extraction, Find Similar Files, and Agent. HTTP Bearer (JWT / Personal Access Token) auth; base URL is your per-tenant Elise domain.
- [API documentation](https://api-docs.biolevatecloud.com/biolevateapi/intro): Human docs for the Biolevate API.

## SDKs & tools
- [Python SDK `biolevate`](https://pypi.org/project/biolevate/): High-level async Python SDK (pip install biolevate).
- [Low-level client `biolevate-client`](https://pypi.org/project/biolevate-client/): OpenAPI-generated client the SDK is built on.
- [onizuka-cli (`ozk`)](https://github.com/Biolevate/homebrew-tap): CLI, install via `brew tap Biolevate/tap && brew install onizuka-cli`.
- [SDK source & changelog](https://github.com/Biolevate/biolevate-api-sdk): Public SDK repo, OpenAPI spec, and release-please changelog.

## Conventions
- Auth: `Authorization: Bearer <pat>`. Idempotency: optional `Idempotency-Key` header on writes. Pagination: page/pageSize (data/totalPages/totalElements/hasNext) and cursor (items/nextCursor). Jobs are async: create → poll GET .../jobs/{jobId} until SUCCESS/FAILED → GET .../results. Ids are UUIDs; timestamps are epoch-millis. Errors are application/json (not RFC 9457).

## Company
- [Biolevate website](https://www.biolevate.com/): Company site.
- [Who we are](https://www.biolevate.com/pages/whoweare)
- [Careers](https://www.biolevate.com/pages/careers)
- [Contact](https://www.biolevate.com/pages/contactus)
- [Blog (Medium)](https://medium.com/@biolevate)
- [LinkedIn](https://www.linkedin.com/company/biolevate)