About

What AnalyzeMyDocument is -- and is not

A structured AI document analysis platform for engineers, researchers, lawyers, and business professionals who need more than a single-pass AI summary.

What AnalyzeMyDocument is

A multi-agent analysis pipeline

AnalyzeMyDocument runs your document through a structured pipeline: a Writer agent produces the initial analysis, three independent Reviewer agents critique it, and a Reconciler agent scores the output and approves or sends it back for revision. This is not a chatbot. It is a pipeline with distinct stages, each with a specific role.

27 specialized analysis types

Each analysis type is a purpose-built prompt template designed for a specific document domain: technical specifications, patents, contracts, research papers, business cases, and more. The pipeline behavior adapts based on the type you select, so the analysis is relevant to your document and your decision.

Confidence-scored output

Every run produces a numeric confidence score (0.0 to 1.0) based on the Reconciler's assessment of completeness, accuracy, and Reviewer consensus. You set the minimum threshold. The pipeline iterates until it meets the bar or exhausts its cycles.

Two modes of operation

Revise Document takes your draft and produces a stronger version. Generate Analysis creates a new analytical document that critiques the original. Both modes use the full pipeline.

What AnalyzeMyDocument is not

Not a chatbot

There is no conversational interface. You submit a document, configure the analysis, and receive a structured output. The pipeline does the work without back-and-forth prompting.

Not a replacement for human experts

The pipeline produces analysis that can inform decisions, but it does not replace the judgment of qualified professionals. Legal, financial, medical, and academic decisions require human expert review.

Not a single-pass summary

Unlike tools that send your document to an LLM once and return whatever comes back, AnalyzeMyDocument runs multiple agents with different perspectives and iterates until the output meets a quality bar.

How the pipeline works

Architecture

AnalyzeMyDocument is a .NET/Blazor application with a stateless LLM integration layer. Document content and pipeline state are stored in SQLite. Analysis artifacts (the generated output files) are stored outside the database as files. The pipeline orchestrator manages the state machine transitions: Pending, Writing, Reviewing, Reconciling, Cycling, Approved, Failed.

Agent execution

Each pipeline stage (Writer, Reviewer, Reconciler) is an independent LLM call with a purpose-built system prompt. The Writer receives the original document and analysis type configuration. Reviewers receive the Writer's output and the original document but not each other's output. The Reconciler receives all Reviewer reports and produces the final assessment.

LLM provider

The pipeline is powered by Claude AI from Anthropic. The application layer uses a provider abstraction (IAgentExecutionService) that decouples the pipeline logic from the specific LLM implementation. Claude-specific details are confined to the infrastructure layer.

Iteration and convergence

When the Reconciler's confidence score falls below the configured threshold, it generates structured feedback identifying what needs improvement. The Writer receives this feedback and produces a revised draft. The full Review and Reconcile stages run again. This continues until the threshold is met or the maximum cycle count is reached.

Professional and academic disclaimer

AnalyzeMyDocument provides AI-generated analysis for informational purposes only. The output produced by this platform does not constitute professional legal, financial, medical, engineering, or academic advice.

Users are responsible for independently verifying all facts, citations, calculations, and conclusions before relying on or submitting the output. The platform does not guarantee the accuracy, completeness, or fitness for purpose of any analysis produced.

Analysis output should be treated as a starting point for further review, not as a final determination. Critical decisions -- including but not limited to legal filings, financial commitments, medical treatment, patent prosecution, and academic submissions -- must be reviewed by qualified human professionals before action is taken.

AnalyzeMyDocument is not liable for decisions made based on the platform's output. By using the service, you acknowledge that AI-generated analysis has inherent limitations and that final responsibility for any use of the output rests with you.

Data privacy

Your documents

Documents you upload are processed through the pipeline and stored only for the duration needed to complete and deliver the analysis. You can delete your documents and all associated output at any time from your account. We do not use your documents to train models.

LLM processing

Document content is sent to Anthropic's Claude API for processing. Anthropic's data usage policies apply to content processed through their API. We use Anthropic's API configuration that opts out of model training on customer data.

Data retention

Completed analysis output is retained in your account until you delete it. Pipeline intermediate artifacts are retained for 30 days after run completion, then automatically purged. API customers can configure custom retention schedules — from immediate deletion after download to extended retention for compliance.

Encryption

All data is encrypted in transit (TLS 1.3) and at rest (AES-256). API keys are stored hashed. We do not log document content in application logs.

Infrastructure

All data — documents, pipeline output, and account information — is stored in Microsoft Azure. Database hosted on Azure SQL. Document storage on Azure Blob Storage. All infrastructure deployed within your selected region (US, EU, or Asia-Pacific).

NDA available

For customers handling sensitive or proprietary documents, we offer a signed mutual NDA as part of a 12-month service agreement. Available to all API tier customers at no additional cost. Request an NDA.

Available everywhere you work

Web Application

Access AnalyzeMyDocument from any modern browser at app.analyzemydocument.com. Upload documents, configure analysis, monitor pipeline progress, and download results. No installation required.

Desktop App (Mac and Windows)

Native desktop application for macOS and Windows. Drag-and-drop document upload, offline queue management, and system notifications when pipeline runs complete. Built with .NET MAUI.

Self-Hosted (Linux)

Run the full pipeline on your own infrastructure. Docker-based deployment for Linux servers. Bring your own Anthropic API key. All document data stays within your network. Ideal for organizations with strict data residency requirements.

REST API

Integrate document analysis into your existing workflows. Submit documents, configure reviews, and retrieve results programmatically. Available on Standard and Deep Analysis tiers. View API documentation.

Try it yourself

Upload a document and see what the pipeline produces. Your first run is free.