v0.1 endpoints available to alpha testers:

• Pre-processed Documents: Access all of our document inputs across 3000 tickers in UK, Europe, US and 3 years of history (50k+ docs)

  • Formats available:

    • Raw HTML, PDF

    • Text-only, text & parsed tables, parsed tables as CSV

    • Markdown

  • Document types:

    • Fillings, Press releases

    • Earnings call transcripts, audio

    • News
      
      

• LLM pipeline outputs: Access all stage of our LLM pipeline programatically including but not limited to:

  • Validation - Document title, document type, content type

  • Description - LLM derived company description driven by recent fillings

  • Retrieved Sentences - Tagged paragraphs from Documents with embeddings and metadata

  • KQuery - Insight 'blocks' per document or period across areas like Revenue, Margins, Management, Strategy

  • KSummary - Final summary blocks built from KQuerys per document or per period

  • Primer - kQuarter summary with reasoned opinion on underlying trends

    All LLM pipeline outputs with associated citations (Retrieved Sentence associations).