Every earnings call, blog, and podcast, indexed by insight, not by keyword.
AI, markets, and technology are moving faster than any person can track. Don't burn your agent's context window searching the open web. Expert System continuously processes hundreds of earnings calls, technical blogs, podcasts, and expert commentary, so your agent stays current and finds the right insight in one query, not ten pages.

Semantic search across thousands of pre-extracted insights. Each result is 2-3 sentences, enough to decide if it's relevant without burning context.
GET /api/v1/takeaways/search ?query=AI+infrastructure+spending → 3 results, ~200 tokens total
Found something relevant? Pull the full earnings transcript, blog post, or research note with metadata and provenance.
GET /api/v1/documents ?ids=doc_xyz789 → Full article text + source link
Ask natural language questions about macroeconomic indicators or company financials. Get clean JSON back, no scraping, no parsing.
POST /api/v1/query/macro
{"query": "Current GDP growth"}
→ Structured FRED dataClaude Code
/plugin install github://starmode-base/expert-system-plugin
Any agent (Cursor, Copilot, Cline, Codex)
npx skills add starmode-base/expert-system-plugin
Works with any agent that supports the Agent Skills standard. Or use the API directly.