Renaissance
Doubt-Driven Research Assistant

The search engine that
doubts itself.

Most AI search engines suffer from confirmation bias—they only find evidence that supports your query. Renaissance is different: it researches your topic, writes a draft, conducts an adversarial critique on its own work, and hunts for the contrarian facts other AIs miss.

State your inquiry
Inquiry Catalysts (Click to investigate)
Case Study in Action

The Self-Skeptical Method

Watch how Renaissance cross-examines its own research. Select a step below to see how it resolves confirmation bias on the question: "Is dark chocolate actually beneficial for cardiovascular health?"

Standard AI Output (Confirmatory Bias)
Hypothesis: Positive

"Yes, dark chocolate improves heart health by lowering blood pressure due to flavanols."

Dark chocolate is rich in flavanols (especially epicatechin), which stimulate nitric oxide production in blood vessels, causing vasodilation and reduced blood pressure. Dozens of published randomized trials confirm flavanols reduce cardiovascular risk.

⚠️ Flaw: Immediately accepts positive findings; ignores study duration, funding source, and caloric offset.
Direct Comparison

Why Renaissance?

Unlike traditional search bots that fetch pages matching your query terms, Renaissance uses an adversarial loop built to challenge assumptions.

ChatGPT + Deep Research

  • Confirmation Bias: Focuses search and summaries on proving your query statement correct.
  • Consensus Fixation: Ignores conflicting study data in favor of writing a smooth, high-level summary.
  • Absolute Certainty: Writes reports in an authoritative tone, hiding gaps and unknowns.

Renaissance

  • Self-Skeptical Search: Specifically designs search queries to look for contrarian facts and counterarguments.
  • Adversarial Audit: Runs an independent critic loop to call out logical leaps and funding conflicts.
  • Calibrated Confidence: Formulates what is strictly known vs. unknown and states exactly what triggers would change its mind.
Cognitive Pipeline Spec

Anatomy of an Investigation

Traditional AI query pipelines generate naive answers immediately. Renaissance executes an iterative, adversarial cycle to assure data precision.

01

Planner

(Deconstruct)

The Planner agent dissects your topic into multiple target inquiries targeting historical, scientific, and logical angles.

02

Crawler

(Deep Scrape)

Runs concurrent web searches and crawls full target web pages, filtering text into high-density semantic chunks.

03

Writer

(Synthesis)

The Writer agent compiles the extracted facts, logs the bibliographic reference list, and drafts the initial thesis document.

04

Critic

(Adversarial Audit)

A Critic agent evaluates the draft, searches for cognitive biases, isolates unverified claims, and suggests counter-queries.

05

Refiner

(Refinement Loop)

Executes secondary scraping runs to fill gaps identified by the Critic before outputting the final verified analysis.

AGENTS_DESK_SIMULATION://PLANNER
ACTIVE MONITOR
Query: Clinical trialsQuery: MechanismQuery: Meta-analyses
>Planner initialised.
>Analyzing topic complexity...
>Generating 3 target inquiry routes:
>-> [Route 1]: Live empirical studies
>-> [Route 2]: Mechanistic/biochemical pathways
>-> [Route 3]: Meta-analyses consensus
>Planner tasks dispatched.
>