Feedback That Explains WHY
Completed professional editor training, five writing courses, earned a qualification in print and digital media. This knowledge just sat on a drive. Ran an experiment: can this knowledge be turned into a working system?
Before / After
Before: Grammarly flags passive voice → fixes it → I accept → I don’t understand why → keep making the same mistake
After: WriteRAG says, “You wrote ‘mistakes were made’ — passive voice. Rule: Strunk & White, Element 14: Use the active voice. Why: Passive hides responsibility.” → I actually understand the principle.
Impact: 85% accuracy in detecting violations. 18 rule categories from 10+ books. Mastered “show, don’t tell” after seven repetitions.
How It Works
Step 1: Extracted rules from 10+ professional writing books, organized into 18 categories.
Step 2: Send text to Telegram bot. AI finds which rules apply to your text.
Step 3: Returns: “Rule violated: Use active voice (Strunk & White, p.18). Problem: ‘was thrown by’ is passive. Why it matters: Active is more direct.” Shows good/bad versions.
Result: Learn writing principles from professional sources, not generic AI advice. Track patterns over time.
18 Rule Categories
Primary: Clarity & concision, Structure & flow, Voice & tone, Show don’t tell, Pacing, Active vs passive
Secondary: Dialogue, Character development, Conflict & tension, Description, Opening hooks, Endings, and 6 more specialized categories
Technical Architecture
- Knowledge Base: 10+ books (On Writing Well, Bird by Bird, Elements of Style, Save the Cat)
- Vector DB (Qdrant): Embeddings for each rule, semantic matching
- Analysis Pipeline: Text → embedding → similarity search → Claude validates → ranked issues
- Learning Loop: Tracks violation frequency, builds “mistake profile”, prioritizes feedback
What Makes It Different
Every piece of feedback cites a real book page or course module. Not AI “opinions,” but professional writing wisdom. It builds intuition through repetition, not dependency.