Matter context.
Bring approved matters and legal data into Claude or any compatible MCP client.
CLI & MCP
One workspace, from every tool.
Search, read, and work with stella from the command line or any MCP-compatible client. Both surfaces share the same capability registry, permissions, and workspace data.
>
Ran 1 CLI command
search_across_matters18 documents
Matched passages9 hits
Printed resultstext · json
Found 9 passages across 18 documents
· Completed in 0.8s · 9 results
▶ stella CLIauthenticated·read-only
In short
What are the stella CLI and MCP server?
They are two interfaces to the same stella capabilities. The CLI mirrors the MCP tool registry as commands, while the MCP server lets compatible AI agents and tools work with matters, documents, templates, and public legal data under the same access controls.
Capabilities
01
Work from the terminal
Use generated commands such as stella search matters, document read, and template fill from scripts or your terminal.
02
MCP-compatible
Connect stella to MCP-compatible clients without building a separate integration for each one.
03
One capability registry
The CLI command surface is generated from the MCP tool registry, so the two interfaces stay aligned as stella evolves.
04
Workspace permissions apply
Authentication scopes and workspace permissions remain the boundary, whichever interface you use.
05
Structured output
Commands return machine-readable JSON for scripts and AI agents as well as readable terminal output for people.
06
Self-host friendly
Point the CLI and MCP clients at the stella server your organisation operates.
01
Search the whole matter from your terminal
- Find passages across matter documents with stella search matters
- Read documents, list matters, and inspect templates without switching context
- Use paginated, structured results in scripts and automated workflows
02
The same capabilities through MCP
- Expose matters, documents, case law, templates, and other approved tools
- Keep the capability catalogue and CLI commands generated from one registry
- Connect compatible AI agents without copying legal data into another workspace
Connect an AI agent
Point any MCP-compatible client at the endpoint below, or at your own server's /mcp path when self-hosting. The first connection opens a browser sign-in, and the agent holds only the scopes you approve.
https://api.stll.app/mcpChatGPT
Turn on developer mode under Settings → Connectors → Advanced, then add a connector with the endpoint address.
Claude
In Settings → Connectors, add a custom connector with the endpoint address.
Codex
Add the server and sign in:
codex mcp add stella --url https://api.stll.app/mcp
codex mcp login stellaClaude Code
Add the server with one command:
claude mcp add --transport http stella https://api.stll.app/mcpAny client that speaks Streamable HTTP with OAuth 2.1 connects the same way. Prefer the terminal? Install the CLI and sign in:
npm i -g @stll/cli
stella auth loginWorkspace, meet MCP. Bring live matter context into the tools where you work.
Tools for AI agents.
Search, read, and fill templates through the same registry that powers the stella CLI.
Only the access you grant.
Authenticate with OAuth; every call remains constrained by granted scopes and workspace permissions.
FAQ
Is the CLI a separate product from the MCP server?
No. They are separate interfaces generated from the same capability registry and connected to the same stella server.
Can I use the CLI in scripts?
Yes. The CLI supports structured JSON output, pagination, predictable exit codes, and non-interactive use suitable for automation.
Does connecting an MCP client bypass workspace permissions?
No. The server checks authentication scopes and stella permissions before a capability runs.