Prism scores your existing metadata against 11 CRO dimensions, then rewrites titles, H1s, and meta descriptions using your brand config — at scale, from your Screaming Frog export.
Prism slots into your existing workflow — import a Screaming Frog export, and Prism does the rest. Two passes: a fast scoring pass that tells you what's wrong, then a quality rewriting pass that fixes it.
Drag in your crawl export. Prism auto-detects columns — URL, Title, H1, Meta Description. Optionally add a tracked keywords CSV for keyword-informed rewrites.
Choose the active client config (brand, persona, tone, USPs, CRO weights). Select which fields to generate — titles, H1s, meta descriptions, or any combination.
Pass 1 scores existing metadata against 11 CRO dimensions — fast and cheap. Review scores before committing to Pass 2, which rewrites using a quality model.
Review rewrites in-app with inline editing, live character counts, and per-row revert. Export as CSV — original and rewritten columns side by side.
Every row is scored against 11 CRO dimensions using a lightweight model. You see the scores before Pass 2 runs — so you know what's being fixed and why.
The quality model receives the CRO scores from Pass 1 plus your client config — brand identity, persona, tone, USPs, and dimension weights. Rewrites are targeted, not generic.
Prism is built around the meta rewriting workflow. Each feature exists because the workflow needed it.
Built on the MECLABS framework. Value, Relevance, Clarity, Urgency, Motivation, Friction, Anxiety and more — scored per row before rewriting begins.
One profile per client — brand identity, customer persona, tone of voice, USPs, and per-dimension CRO weights. Import/export as JSON. Existing AppScript CRO Master V7 files import without conversion.
Three-state colour coding per field. Green (within target), amber (marginally off), red (significantly out of range). Counts update live as you edit inline. No hard truncation.
LLM calls go directly from Prism to OpenAI, Gemini, or Anthropic using your own API keys. No Search Labs middleware. No per-rewrite charge on top. API keys stored in Windows Credential Manager.
All client data, profiles, and outputs stay on your machine. No Search Labs server ever sees your client metadata. Built on Tauri v2 — ~5MB binary, no Electron runtime bloat.
Step-by-step wizard that builds a client config profile through structured questions, with Gemini deep research integration to pre-populate brand and competitor context automatically.
The meta rewriting pipeline is built and working. Register your interest and we'll notify you when Prism is available for download — along with early access pricing for registered users.
