WriterzRoom runs three generation pipelines. Each uses different AI models, different agent counts, and different levels of research depth. Your plan determines which tiers you can access. Your choice of tier at generation time determines the credit cost and output quality.
Quick is a 3-agent pipeline optimized for speed. It runs the planner, writer, and a lightweight formatting pass using Claude Haiku across all agents. No research agent, no editor, no SEO optimization.
Use Quick for first drafts, internal documents, brainstorming output, or any content where speed matters more than polish. Quick generations typically complete in 15-30 seconds.
Credit cost: 1 credit per generation.
Agents: Planner → Writer → Formatter
Model: Claude Haiku 4.5 (all agents)
Standard
Standard is the full 7-agent pipeline. It adds the researcher (Tavily-powered web search), a dedicated call writer coordinator, a full editor with readability enforcement and AI-tell detection, SEO optimization, and a publisher agent. The writer and editor use Claude Sonnet 4 for higher output quality. All other agents use Claude Haiku 4.5.
Use Standard for published content, client deliverables, blog articles, press releases, and any content that needs to read well and be factually grounded. Standard generations typically complete in 60-120 seconds depending on research depth.
Credit cost: 4 credits per generation.
Agents: Planner → Researcher → Call Writer → Writer → Editor → Formatter → SEO → Publisher
Models: Claude Haiku 4.5 (planner, researcher, formatter, SEO, publisher), Claude Sonnet 4 (writer, editor)
Research: Up to 3 Tavily search queries per generation, with source verification and inline citation.
Quality gates: Readability enforcement, AI-tell detection (max 5), word count validation with up to 2 expansion attempts.
Premium
Premium uses the same 7-agent pipeline as Standard but upgrades the writer to Claude Opus 4. It also includes human-in-the-loop review at the editor stage, where the editor pauses and presents suggestions before the content moves to formatting. Premium research runs up to 5 search queries.
Use Premium for high-stakes content: investor presentations, research papers, white papers, competitive analyses, and any content where maximum quality justifies the cost and the additional review step.
Credit cost: 5 credits per generation.
Agents: Planner → Researcher → Call Writer → Writer → Editor (with HITL pause) → Formatter → SEO → Publisher
Models: Claude Haiku 4.5 (planner, researcher, formatter, SEO, publisher), Claude Opus 4 (writer), Claude Sonnet 4 (editor)
Research: Up to 5 Tavily search queries per generation.
Quality gates: Same as Standard, plus human-in-the-loop editor review and up to 3 expansion attempts.
Choosing the Right Tier
| Consideration | Quick | Standard | Premium |
|---|
| Time to complete | 15-30s | 60-120s | 90-180s + review |
| Research included | No | Yes (3 queries) | Yes (5 queries) |
| AI-tell suppression | Prompt-level only | Prompt + editor gate | Prompt + editor gate |
| Editor review | None | Automated | Human-in-the-loop |
| Expansion attempts | 0 | Up to 2 | Up to 3 |
| Best for | Drafts, internal docs | Published content | High-stakes deliverables |
Start with Standard for most published content. Use Quick for iteration and brainstorming. Reserve Premium for content that justifies the extra cost and review time.
Plan Access
| Plan | Quick | Standard | Premium |
|---|
| Demo | Yes | — | — |
| Starter | Yes | — | — |
| Professional | Yes | Yes | — |
| Enterprise | Yes | Yes | Yes |
