# Ellora > Your editor in chief, for every AI draft. Your team already writes with ChatGPT and Claude. We built Ellora to be the AI content editor inside those apps. It learns how your brand sounds, then fixes each draft before anyone has to. Categories: AI content editor, brand voice, editorial review for AI drafts. ## Links - [Agent skill (SKILL.md)](https://getellora.com/SKILL.md): install, OAuth, and the review loop - [Docs](https://getellora.com/docs): HTML view of the same SKILL.md source - [Security](https://getellora.com/security): https://getellora.com/privacy and https://getellora.com/terms - [Support](https://getellora.com/support): ask us directly at hello@getellora.com - [Full agent facts](https://getellora.com/llms-full.txt): plans, privacy, and setup detail - [MCP server card](https://getellora.com/.well-known/mcp.json): discovery for remote MCP clients - [Create account](https://app.getellora.com/register): Free plan, one writer - [Sign in](https://app.getellora.com/login) ## What Ellora is Ellora is a hosted MCP server that reviews drafts in the LLM apps your team already uses. It runs **VoiceDNA** (voice conformance from your published writing), **AI Editor** (desloping), **Narrative**, and **Editorial** (writing brief and dictionary rules) on each draft before it ships. Ellora has been on the [official MCP Registry](https://registry.modelcontextprotocol.io) since 19 October 2026. ## How it works 1. Add the remote MCP server and authenticate with OAuth (no API key). 2. Build a voice profile from 2–20 samples of your team's writing. 3. Set a writing brief per format (audience, intent, format, vocabulary). 4. On each draft, run the `ellora_review` loop until the session reports `verification.passed: true` or `diminishing_returns`. Give your app the skill at https://getellora.com/SKILL.md once; after setup, writers keep working in their usual app. ## Team value Teams use Ellora for shared standards: one voice profile per brand or author, briefs and the dictionary shared across members, review history with roles (admins control who joins), and a dashboard of words reviewed — so every draft meets the same bar without a separate copy desk for each app. ## Plans - **Free:** one writer, one voice profile, every check on every draft. Create an account at https://app.getellora.com/register. - **Team:** more voice profiles and team members. Email us at hello@getellora.com or set up a call at https://cal.com/getellorademo/client-demo. ## Works in Claude, ChatGPT, Cursor, Perplexity, Grok, OpenCode, and any client that supports remote MCP over Streamable HTTP. ## Install (config-file clients) Add this entry to the client's MCP settings (Cline-compatible `streamableHttp` type), then restart and complete OAuth: ```json { "mcpServers": { "ellora": { "type": "streamableHttp", "url": "https://api.getellora.com/v1/external/mcp" } } } ``` OpenCode uses its own `mcp` block: ```json { "$schema": "https://opencode.ai/config.json", "mcp": { "ellora": { "type": "remote", "url": "https://api.getellora.com/v1/external/mcp" } } } ``` Then run `opencode mcp auth ellora`. OpenCode can also install the skill with `npx skills add getellora/agent-skills -a opencode`. CLI: `claude mcp add --transport http ellora https://api.getellora.com/v1/external/mcp` (Claude Code) or `cmd mcp add --transport http ellora https://api.getellora.com/v1/external/mcp` (Command Code), then authenticate. OAuth metadata: https://api.getellora.com/.well-known/oauth-protected-resource and https://api.getellora.com/.well-known/oauth-authorization-server. ## Who it's for In-house content and brand teams, product marketing, and agencies whose writers draft in ChatGPT, Claude, Cursor, or similar — anyone who needs one editorial standard across many AI drafts. ## Privacy (summary) Drafts and reviews stay in your workspace; nothing is shared across customers; your content is not used to train models. Full wording: https://getellora.com/privacy and the "Privacy" section in https://getellora.com/llms-full.txt.