Launch Directory/AI & ML/XERJ Reader-1 — neurosymbolic data model
XERJ Reader-1 — neurosymbolic data model

XERJ Reader-1 — neurosymbolic data model · Launch Video Breakdown: Hook, Pacing & Motion Design

Answers document questions with verifiable programs that cite exact source words

0:00 · The Hook · Product Announcement & Mission
0:00 / 0:00
AI & MLProductionNovember 12, 2026

Direct Value Proposition & Contrarian Comparison

18.2

Cuts / min

3:25

Runtime

513K

Views

653

Likes

0

Reposts

0

Replies

Scene-by-scene timeline & spoken transcript

  1. The Hook

    Product Announcement & Mission

    “Today, we're announcing our first neurosymbolic model that can actually handle real work with data.”

    On screen
    XERJ Reader A neurosymbolic model for real work with data. Re: your daily RAM report Total installed is 64 GB, and the busiest moment was 21.3 GB
    Camera
    Smooth dolly zoom out revealing scattered documents on a dark canvas, pushing into an email card.
    Motion
    Floating 3D plane physics with dynamic text highlight overlays.
  2. Feature Teaser

    Cost & Efficiency Benchmark

    “Trained specifically for data extraction, Search Reader is tens of times cheaper than even the most budget friendly LLMs, with performance closing in on frontier models.”

    On screen
    PRICE PER QUESTION Short document, 1 question XERJ Reader-1 Claude Haiku 5.5 Claude Opus 5.5 $0.0000065 $0.000172 $0.00688 1x 26x ~1,050x Performance: closing in on frontier models.
    Camera
    Static front-facing dashboard view with sequential vertical column reveals.
    Motion
    Animated number tick-ups and spring-damped column highlight cards.
  3. Problem Agitation

    Paradigm Shift Typography

    “It isn't just better, it is a different type of AI model. Here's how it works.”

    On screen
    IT ISN'T JUST BETTER. A DIFFERENT TYPE OF AI MODEL. HERE'S HOW IT WORKS.
    Camera
    Snapping cut with drifting background document particles.
    Motion
    Kinetic bold sans-serif text transitions with yellow accent callouts.
  4. Product Reveal

    Generative Hallucination vs. Formulaic Precision

    “Frontier LLMs were trained to chat. Ask one about your data and it writes an answer that sounds right, whether it's in the data or not. So a human still has to check it. Reader was trained for exact answers. It points at your data and writes a formula, like in Excel, to get the answer.”

    On screen
    LLM chat How much memory is just sitting there doing nothing? Memory just sitting there doing nothing: 64 GB LLM answer 64 GB Human check Reader fx = SUB(NUM(d1[155:160]), NUM(d1[189:196])) -> 42.7 GB
    Camera
    Pan from simulated chat window to parsed document, then camera tilts into formula bar.
    Motion
    Character-by-character typing emulation, vector connector lines, and UI badge morphs.
  5. Feature Teaser

    Macro Cost and Latency Metrics

    “It is built to dramatically reduce cost and latency...”

    On screen
    COST $6,880.00 $6.50 per 1M questions Claude Opus 5.5 XERJ Reader-1 LATENCY ~1.70 s 0.08 s Claude Opus 5.5 to first token XERJ Reader-1 full answer
    Camera
    Locked off centered split screen comparison.
    Motion
    Strikethrough animations and instant latency scale bar reductions.
  6. Feature Teaser

    Use Case Showcase

    “...within tasks like finding facts and contradictions in large documents, building an enterprise search system, curating context within coding agents, or figuring out how many shares you keep after leaving Anthropic.”

    On screen
    Do the refund rules contradict each other? 96-PAGE POLICY Yes: 30 days vs 14 days Who pays for shipping? COMPANY KNOWLEDGE BASE 2,140 FILES The Supplier Where is the retry limit set? CODING AGENT MAX_RETRIES = 5 How many shares do I keep after leaving? 1,250 vested shares
    Camera
    Rapid sequential fly-ins and zooms into simulated document cards, code files, and email clients.
    Motion
    Text bounding-box highlighter tracks with contextual callout chips.
  7. Product Reveal

    Hybrid Pipeline Architecture

    “Reader does this by taking the data heavy lookups, while the LLM handles workflow steps that are open ended.”

    On screen
    LLM OPEN-ENDED STEPS Plan the task Decide the offer Write the reply XERJ Reader-1 DATA-HEAVY LOOKUPS Find the order Find the clause Find the payment 0.08 s
    Camera
    Top-down birds-eye schematic layout overview.
    Motion
    Card re-sorting animation categorizing process pipeline steps into distinct swimlanes.
  8. Feature Teaser

    Long Document Retrieval

    “It can even answer questions over long documents on its own, like finding a clause in a contract or a due date in an invoice.”

    On screen
    How much notice stops the auto-renewal? PAGE 3 OF 186 PAGE 17 OF 186 PAGE 37 OF 186 written notice of non-renewal not less than 90 days When is invoice INV-21094 due? PAGE 31 OF 31 27 November 2026
    Camera
    Horizontal card conveyor belt scrolling through multidocument stacks.
    Motion
    Page pagination counters and instant yellow bounding box snap.
  9. Feature Teaser

    Side-by-Side Live Benchmark

    “The improvements are clear if you see them side by side. Ask Search Reader a question about your data just like you would an LLM. Get the answer back near instantly. Meanwhile, LLMs take many times longer to finish responding. Look at how the LLM searches through the data, file by file, then generates its answers token by token, which is great for a natural conversation, but slow and costly for computers. Or the LLM writes a script, runs it, reads the output, and fixes it. A single lookup takes several model calls. Reader writes one formula, like in Excel. It points at the words it needs, and your code runs it.”

    On screen
    XERJ Reader-1 Claude Opus 5.5 0.08 s to the exact span 3.09 s to finish the answer CLAUDE OPUS 5.5 - INPUT TOKENS 664 file 6 of 6 output tokens 34 via a script model call 4 SUB(NUM(d1[155:160]), NUM(d1[189:196])) -> 42.7 GB
    Camera
    Split-screen dashboard displaying real-time timers and token counters.
    Motion
    Dual synchronized timers, progressive token rendering vs single-shot instantaneous calculation.
  10. Call to Action

    Closing Paradigm & Waitlist

    “This represents a fundamental shift in how AI agents work with data. Join the waitlist at platform.xerj.ai”

    On screen
    XERJ Reader-1 How much did Sam pay for this RAM? No answer Join the waitlist platform.xerj.ai Found, not generated.
    Camera
    Pull back from graph nodes to centered CTA card with floating ambient pages.
    Motion
    Node network pulse lines resolving into clean centered CTA card.

Related AI & ML Product Launches

Explore all AI & ML launches →
Grok Bot 1.0 — AI teammates for your team
@bot
AI & ML
@bot
Grok Bot 1.0 — AI teammates for your team@bot
Tauvus
@tavus
AI & ML
@tavus
Tauvus@tavus
Fish Audio
@FishAudio
AI & ML
@FishAudio
Fish Audio@FishAudio