Writer Tools

A suite of AI-powered writing assistants for blogs, essays, articles, stories, presentations, and ad copy.

Overview

Writer Tools is a collection of specialised AI writing screens inside TotalApp. Each screen is purpose-built for a specific content type: blog posts, academic essays, news articles, creative stories, anime scripts, slide presentations, ad copy, cinematic scenarios, and text rewriting.

Available Tools

Blog Writer

Multi-stage blog post generation with outline, introduction, body sections, and conclusion.

Essay Writer

Structured academic or argumentative essays with thesis, supporting paragraphs, and citations.

Article Writer

News-style or long-form articles with a lead paragraph, subheadings, and pull quotes.

Storyteller

Creative fiction with characters, plot arcs, dialogue, and genre controls.

Anime Writer

Anime episode scripts with scenes, character lines, and stage directions.

Presentation Wizard

Slide deck outlines and speaker notes generated from a topic or document.

Ad Writer

Short-form ad copy for Google Ads, social media, email subject lines, and CTAs.

Scenario Writer

Cinematic screenplays and scene-by-scene story scenarios.

Smart Rewriter

Rewrites pasted text to be more professional, more fluent, or shorter — without changing its core meaning.

Prompt Library & Knowledge Attach (Semantic Search)

Every Writer Tool screen (except Presentation Wizard) has two buttons above its main text field:

Insert Prompt

Opens your saved prompt library (My Prompts) and inserts a reusable prompt template into the field — useful for house style guides, recurring instructions, or brand voice reminders.

Attach Knowledge

Opens a picker over your personal library (My Knowledge — uploaded files, URLs, and pasted text) so generation can be grounded in real reference material instead of the model's general knowledge.

Semantic Ranking

As you type, attached knowledge sources are ranked by embedding-based relevance to your input, with a relevance score badge shown in the picker. Configure the embedding model in Settings → Agentic → Embedding Engine.

Assisted vs. Automatic

Assisted (default): you manually confirm which ranked sources to attach. Automatic: the top-ranked sources auto-attach as you type — configure the mode in Settings → Agentic → My Knowledge Search Behavior.

Without an embedding model configured, Attach Knowledge still works — the picker simply shows your library unranked, and generation is never blocked.

AI Engine

All Writer Tools use the engine selected in Settings → Doc View → Writer Engine:

API (default): Anthropic Claude or Simplified.com API depending on the tool
Local CLI: runs Claude via the Claude Code CLI on your machine
Local LLM: BERT-based local model (intent classification only; generation not supported)
Ollama: runs a locally installed Ollama model on your own machine for text generation, fully offline

Semantic search (Attach Knowledge ranking and Automatic auto-attach) is configured separately in Settings → Agentic → Embedding Engine and currently supports Ollama embedding models — independent of which Writer Engine you have selected above.

Frequently Asked Questions

Can I save and export generated content?

Yes. Every generated piece can be saved to My Documents and exported as DOCX, PDF, or plain text.

Do all tools support multiple languages?

Writer Tools generate content in the language of your prompt. Write your topic in Turkish to get Turkish output, English for English, and so on.

What's the difference between Insert Prompt and Attach Knowledge?

Insert Prompt pastes a short reusable instruction or style template into your input field. Attach Knowledge injects the full content of one or more library sources as context for the AI — for citing facts, referencing a document, or grounding generation in something you already wrote.

Can I generate content fully offline?

Yes. Set Settings → Doc View → Writer Engine to Ollama to run a locally installed Ollama model for generation on your own machine, with no data sent to an external API. Note that the Local LLM (BERT) option only supports intent classification, not text generation.