Free AI-powered California legal research.
Type your question in plain English. The system extracts the right filters, searches both keyword + semantic, and can synthesize a cited answer. No login. No paywall. Ever.
Why this exists
Every California Superior Court publishes tentative rulings — preliminary decisions a judge intends to issue before the actual hearing. They're public records, but they live scattered across dozens of court websites in inconsistent formats, often disappearing within days. There's no single place to search across them, and the existing paid legal-research tools either don't index this data at all or charge enterprise pricing for it.
We built this because we needed it ourselves as working attorneys, then decided to make it available to everyone. The whole stack is engineered to run on permanent free tiers so usage growth never turns into per-user cost.
What you can do
Natural-language search
Type "summary judgment denied in employment cases this month" — the system extracts filters, runs hybrid keyword + semantic search, ranks via Reciprocal Rank Fusion.
Cited LLM answers
Ask research questions and get a 2-3 paragraph answer with inline citations to the source rulings.
Ask about any ruling
Every case detail page has a streaming Q&A widget grounded in just that ruling. "Why did the court rule this way?" — get the answer in 3 seconds.
Cited-authority extraction
Every ruling automatically extracts the cases, statutes, and rules it cites — grouped, normalized, jurisdiction-tagged.
Judge dossiers + comparison
Per-judge grant rates, motion-type breakdowns, plaintiff-vs-defendant splits. Compare two judges side-by-side.
Email digests
Sign up for a daily or weekly email with new rulings matching your saved query. One-click unsubscribe in every email.
RSS feeds
Per-court, per-judge, and global feeds. Subscribe in your feed reader and never poll a court website again.
100% free
No login, no paywall, no premium tier. Pages are CDN-cached; LLM features run on free-tier Gemini credits.
How it works under the hood
A GitHub Actions scraper runs every 15 minutes during California business hours, pulling new PDFs from each court's public website (defeating Cloudflare Turnstile via CapSolver for the holdouts). Each PDF is split per case by Gemini, structured into XML, and embedded into a Postgres + pgvector database. The web app reads from there.
The pipeline is fully open-source: the scraper repo publishes XMLs to GitHub; this web app reads them via a webhook on every successful scrape. Architecture details: ARCHITECTURE.md
Get the daily digest
Built by a California attorney for California attorneys. Source on GitHub. This service is for informational use only and is not legal advice.