Skip to content

RoastMyPost

📋Page Status
Page Type:ResponseStyle Guide →Intervention/response page
Quality:35 (Draft)⚠️
Importance:25 (Peripheral)
Last edited:2026-02-01 (5 days ago)
Words:690
Structure:
📊 5📈 0🔗 8📚 418%Score: 11/15
LLM Summary:RoastMyPost is an LLM tool (Claude Sonnet 4.5 + Perplexity) that evaluates written content through multiple specialized AI agents—fact-checking, logical fallacy detection, math verification, and more. Aimed at improving epistemic quality of research posts, particularly in EA/rationalist communities. Significant false positive rate means it's a complement to, not replacement for, human review.
Issues (1):
  • QualityRated 35 but structure suggests 73 (underrated by 38 points)
DimensionAssessmentEvidence
InnovationModerateMulti-agent evaluation approach for document review
Practical ImpactGrowingUseful for pre-publication review of research posts
Technical MaturityExperimentalDeveloper acknowledges significant false positive rate
IntegrationGoodDirect import from LessWrong and EA Forum
AccessibilityHighFree, web-based, no setup required
Output QualityMixedHelpful for catching errors but requires human filtering
AttributeDetails
NameRoastMyPost
OrganizationQURI (Quantified Uncertainty Research Institute)
LeadOzzie Gooen
LaunchedDecember 2025
Primary ModelClaude Sonnet 4.5
Fact-CheckingPerplexity integration
Websiteroastmypost.org
SourceGitHub (open-source)

RoastMyPost is an experimental web application that uses large language models to evaluate written content through multiple specialized AI evaluators.1 Developed by Ozzie Gooen at QURI, the platform analyzes documents for errors, logical fallacies, factual inaccuracies, and other issues that human reviewers might miss or find tedious to check manually.

The tool is designed to provide “roasts” — critical feedback that highlights potential problems in written work before publication. Unlike general-purpose AI assistants, RoastMyPost deploys specialized evaluator agents that each focus on specific types of analysis.

The platform is particularly relevant to the AI safety and rationalist communities, as it can import posts directly from LessWrong and the EA Forum via URL, making it easy to get feedback on research posts common in these communities.

  • Direct text: Paste markdown content directly
  • Forum URLs: Import posts from LessWrong and EA Forum automatically
  • Web URLs: Extract content from general web pages

RoastMyPost runs multiple specialized evaluators in parallel:1

EvaluatorFunction
Fact CheckerUses Perplexity searches to verify factual claims
Spelling/GrammarIdentifies language errors
Logical Fallacy DetectorFlags potential reasoning errors
Math VerifierChecks mathematical equations and calculations
Link ValidatorTests whether referenced URLs are accessible
Binary Forecast CheckerCompares predictions against actual outcomes
Epistemic AuditorHigh-level assessment of reasoning quality

Processing typically completes in 1-5 minutes depending on document length.

  • Inline annotations: Specific comments highlighted in the text with importance ratings
  • Summary reports: Overall assessment and key findings
  • Grades: Letter grades for different quality dimensions
  • Export: XML export for further processing

Works best with:

  • Documents between 200-10,000 words
  • Content containing factual claims that can be verified
  • Research posts and analyses
  • Squiggle probabilistic models

Less suitable for:

  • Very long documents (performance issues)
  • LaTeX-formatted content
  • Highly specialized technical content requiring domain expertise

The developers explicitly acknowledge significant limitations:1

LimitationDescription
False positivesSignificant rate of incorrect error flagging
Context gapsLacks nuanced understanding for some interpretations
Fallacy checkerSometimes flags valid reasoning patterns
Complex fact-checkingStruggles with claims requiring multiple research iterations
No domain expertiseCannot replace human expert review in specialized fields

The platform is experimental and should be used as one input among many rather than a definitive quality assessment.

Ozzie Gooen has committed to dedicating approximately one-third of his annual work time to maintaining and improving RoastMyPost.1 The roadmap includes model updates as new Claude versions become available and improved evaluator accuracy.

RoastMyPost is currently free for reasonable use, funded through QURI. Usage limits exist to prevent abuse.

ToolPurposeRelationship
SquiggleProbabilistic modeling languageRoastMyPost can evaluate Squiggle models
SquiggleAILLM model generationShared LLM integration patterns
ElicitResearch assistantSimilar LLM-for-research space
  1. Announcing RoastMyPost: LLMs eval blog posts and more, EA Forum, December 2025 2 3 4