How TextScore Works

TextScore analyzes your text in two stages. The first stage runs instantly on our backend. The second stage activates only when the first stage finds signals that need deeper review. This two-phase design keeps results fast for most content while still catching nuanced risks when they appear.

Phase 1 runs entirely on our servers with no external API calls. It checks your text against 12+ metrics - readability, passive voice, link security, spam triggers, punctuation patterns, capitalization, sentiment, and more. Results arrive in under 200 milliseconds.

What Phase 1 Covers

• Speed: Under 200ms per analysis. No loading spinners, no waiting.
• No external calls: Your text never leaves our backend during this phase. No third-party APIs involved.
• 12+ metrics: Readability, passive voice, link security, spam patterns, forbidden keywords, content quality, sentiment, punctuation, AI writing patterns, capitalization, professionalism, and word count.
• Always available: Phase 1 runs on every check, regardless of your plan or usage level.

How It Works Technically

Phase 1 uses pattern matching, statistical formulas, and dictionary lookups. The Flesch Reading Ease formula scores your readability. Regex patterns catch spam punctuation and capitalization abuse. A curated keyword database flags restricted terms. Each metric returns a score and a tier (good, fair, or poor) that feeds into your overall report.

Phase 2 activates conditionally. If Phase 1 detects moderate risk or finds signals that need context to interpret, the system sends your text to an AI model for deeper review. This phase provides qualitative feedback that pattern matching alone cannot produce.

What Phase 2 Adds

• Contextual understanding: The AI reads your text as a whole, not just as isolated patterns. It understands sarcasm, nuance, and intent.
• Shadowban risk assessment: Based on the full context, the AI estimates how likely your content is to be suppressed on your target platform.
• Writing improvement suggestions: The AI Writing Coach gives you specific, actionable rewrites - not generic advice.
• Platform-specific insights: Recommendations tailored to where you plan to post (X, LinkedIn, Medium, Discord, and others).

When Phase 2 Triggers

Phase 2 does not run on every check. It activates when Phase 1 finds a combination of warning signals, when your overall score falls into the fair or poor range, or when specific high-risk patterns appear (like potential shadowban triggers). This keeps costs down and speeds up results for clean content.

Every analysis returns a structured report. Phase 1 results always appear. Phase 2 results appear below them when triggered.

• Overall score: A single number that summarizes how your content performs across all metrics.
• Per-metric breakdown: Each of the 12+ metrics gets its own score, tier (green/yellow/red), and explanation.
• Flagged items: Specific words, phrases, or patterns that triggered warnings, with location references.
• AI feedback (Phase 2 only): Written analysis and rewrite suggestions from the AI Writing Coach.