Elyra
Elyra The coding agent eTerm The terminal that knows where each command ends e The native code editor Elyra Grove Native local development environment Askr The real server for Laravel & PHP Elyra Framework Rust + Svelte 5 framework for desktop apps Elyra Conductor Local project conductor Elyra SQL Server MySQL-compatible SQL server in Rust Elyra Félagi Agents as teammates on one board Elyra SQL Client Native desktop SQL workbench Elyra SQL Anywhere Replication-ready SQL engine Elyra Sjá SEO & GEO workspace for macOS Elyra DataGrid Server-driven data grid for Laravel
Release notes
Changelog
Elyra
Elyra Sjá — Documentation

Elyra Sjá — Documentation

Elyra Sjá (Old Norse sjá, "to see") is a local-first SEO & GEO (Generative Engine Optimization) workspace for Apple Silicon macOS. It shows you what search engines and AI answer-engines see on a page, turns each issue into an optimized fix prompt, and — with your code folder connected — lets an AI agent apply the fixes directly.

Built on the Elyra Framework (Rust + Svelte 5), compiled to a single native binary.

Guides

Doc What it covers
Getting started Install, trial/license, first analysis
Features Analysis, what the model sees, the citation probe, structured data & entity, hreflang, locale, crawl, AI crawlers, generation, CLI, MCP, monitors
Settings AI, citation probe keys, PageSpeed, Search Console, locale defaults, branding
MCP server Use Sjá from Claude Desktop, Zed or Cursor — setup, tools, examples
Licensing Trial, activation, key format for the website
Architecture Modules, commands, database schema
Security & privacy Keychain, local data, the AI sandbox

At a glance

  • Analyze a page → weighted checks across six categories — Meta, Structure, Social, GEO/AI, Indexing, International (+ optional Core Web Vitals) → score & grade. A category that wasn't measured is left out, not scored 100.
  • See what the model sees — the page's main text as an answer engine extracts it, a ≈2 000-token reading budget drawn across it, and the passages most likely to be quoted, each scored and explained. GEO first asks whether the answer crawlers may fetch the page at all, and whether the server lets them.
  • Measure citations — ask ChatGPT, Claude and Perplexity the question a page answers, with your own keys, and see whether they cite it and whom they cite instead, over time.
  • Target a market — pick a language/country and the page is fetched as that market receives it (Accept-Language), with the locale recorded on the report.
  • Crawl a whole site (sitemap + internal links, stoppable mid-run) → aggregated issues, broken pages, duplicates.
  • Track every run in projects & history; watch trends and compare runs.
  • Fix issues with an AI agent that edits your actual files (sandboxed, with confirmation).
  • Prove whether AI crawlers actually visit — from your own server log, with the robots.txt verdict per crawler (GPTBot, OAI-SearchBot, ClaudeBot, …), visits per day over time, and each visit verified against the operator's published address ranges.
  • Generate llms.txt and JSON-LD from crawl data, never invented, saved through the same confirmation as AI edits.
  • Gate a deploy with elyra-sja check <url> --budget 80, or let your own AI assistant call the engine over MCP.
  • Monitor URLs on a schedule with regression alerts.
  • Connect Google Search Console with your own OAuth client — app-wide or one account per project.
  • Report to a branded, client-ready PDF.

Support

Commercial software — see Licensing. Buy or manage a license at https://elyracode.com/sja.