Flagship Product

CLERK Deep Research

Clerk is Ngentech’s flagship deep research agent — a Distributed Durable Execution Engine that breaks topics into 100+ parallel sub-niches, maintains state for up to a week, and delivers statistical analysis that moves beyond “AI guessing” to actual data science.

100+

Parallel sub-niches per topic

1 week

Durable task persistence

4 phases

Research pipeline

0–1

Source reliability scoring

Why Clerk is S-Rank

Four Pillars

Durable

Most agents die if the internet flickers. Clerk is backed by Temporal.io — tasks survive server restarts for 1+ weeks.

Deep

100 sub-niches ensure no stone is left unturned. Clerk searches “Web3 + Zero Knowledge Proofs + 2014”, not just “Web3”.

Quantitative

Moves beyond “AI guessing” to data science — calculating answers based on a mass of personally collected data.

Auditable

Full commit history shows how Clerk’s thoughts and odds evolved over the research period.

The Pipeline

From Topic to S-Rank Report

Phase I

Decomposition

Commander Agent

A topic like “Web3” is broken into 100 specific sub-niches by frontier reasoning models. Output: a Research Map with unique IDs, timeframes, and search parameters.

Phase II

Distributed Extraction

Worker Agents ×100

100 simultaneous Temporal workflows fan out — each researching one sub-niche with neural search, clean scraping, and schema-enforced extraction of reliability scores, technical claims, and sentiment.

Phase III

Aggregation & Statistics

Analyst Agent

All results are loaded into DuckDB. Correlation matrices, trend vectors, and predictive “odds” are calculated with deduplication across sub-niches.

Phase IV

Synthesis

Synthesizer Agent

Claude 3.5 compiles the statistics into an S-Rank enterprise research paper — executive summary, citations, visualizations, and predictive odds analysis.

Multi-Agent System

The Agents

o3 / DeepSeek-R1

Commander

Topic decomposition & research map generation

Claude 3.5 + GPT-4o-mini

Workers

Parallel research, extraction, quality scoring

DuckDB + Polars

Analyst

Statistical analysis, correlations, deduplication

Claude 3.5

Synthesizer

Final S-Rank report, citations, predictive odds

Research that computes, not guesses.

From Web3 to AI agents, see how Clerk decomposes a topic into 100+ sub-niches and delivers statistically-backed odds.