stella, in machine-readable form

stella is an open-source legal workspace and data stack. The workspace covers matters, documents (including Word .docx editing), document review, research, and an AI agent; the data stack adds a case-law reader for official public legal sources and typed access to national company registries such as ARES, Companies House, and SEC EDGAR. It is built for legal teams that want control over their tooling and data: open source, self-hostable, and exportable.

stella includes a case-law reader built on its Legal Atlas ingestion stack for official public legal sources, plus typed clients for national company and commercial registries — including ARES (Czechia), Companies House (UK), SEC EDGAR (US), KRS (Poland), PRH (Finland), and VIES (EU VAT). The AI agent can pull from these sources directly inside a matter, with citations.

The agent chats over your matters, files, connected registries, and tools, with approvals and source previews so you can see where every answer comes from. It also powers Tabular Review, which extracts structured answers from large document sets into matter-scoped review tables for due diligence, discovery, and research.

Yes. stella includes reusable document templates with fields and conditional clauses. You fill a template in by hand or let stella draft it from a matter's data; the conditional clauses resolve, and the assembled document lands in the matter as a draft you keep editing. Clauses are reusable across templates, so repeat documents start from a known-good base.

Yes. The core workspace is free to use, and self-hosting is free under the Apache-2.0 license. You only pay for AI usage, and with BYOK that cost goes directly to your own AI provider.

Yes. stella is open source under Apache-2.0 and can run on your own infrastructure, or you can use the managed cloud. Self-hosting uses standard formats with full export, so there is no lock-in.

stella is designed for privileged legal work: workspace isolation, role-based access, tenant-scoped queries, encryption in transit and at rest, and audit logging for sensitive actions. You can self-host so data stays on your infrastructure, and AI is bring-your-own-key so requests go to the provider you choose.

You choose. Self-host stella wherever you run it, or use the managed cloud. Either way your data is exportable in standard formats at any time.

AI features are bring-your-own-key: requests go to the AI provider you configure, under that provider's terms. Review your provider's retention and data-use settings to control how your data is handled.

Yes. stella supports the Model Context Protocol (MCP) and a CLI, plus anonymization so agents can work with material without exposing names or identifying details. AI outputs are grounded by citations traceable back to the source text.