A Whizz Tech product · Whizz Scribe

Your competitors publish every day. You publish when there's time.

Whizz Scribe is an AI content writer that researches your topic, writes in your brand's voice, and publishes straight to your site — on its own.

I'm a developer — open the tool

3 minutes

from idea to a published, cited article — not 3 weeks

From $29/mo

less than a single agency blog post

Real sources

researches live and cites what it claims — no made-up facts

Changelog

What shipped.

Dated and specific. If it's listed here, it's live.

publishing

Contentful connector + A/B title variants

  • Contentful joins WordPress and Sanity — entries created via the CMA, environment-aware.
  • Every article now ships three title variants with distinct hooks (number, how-to, contrarian) for A/B testing.
  • POST /v1/articles/:id/publish accepts as_draft; the response returns the external URL either way.
dashboard

Dashboard: review queue, usage analytics, credit ledger

  • Article review queue — generated pieces land in in_review with the full quality breakdown attached.
  • Usage analytics per key and per endpoint: credits, model compute, tokens in/out.
  • The credit ledger is now visible line by line, including automatic refunds for failed jobs.
api

Webhooks, idempotency, and the docs site

  • job.succeeded / job.failed webhooks, signed per the Standard Webhooks spec (webhook-id / -timestamp / -signature).
  • Idempotency-Key on POST /v1/articles — retries return the original job instead of double-charging.
  • Docs shipped: quickstart, authentication, full endpoint reference, error and rate-limit tables.
engine

Knowledge base v2: contextual chunks + hybrid search

  • Chunks are contextualized against their source document before embedding — retrieval quality up measurably on multi-page sites.
  • Hybrid (vector + keyword) search across the brand knowledge base during the research stage.
  • Voice profiles now prefer writing samples when present; banned-phrase lists merge into the edit pass.
engine

The engine: seven-stage pipeline with per-stage telemetry

  • Pipeline goes live: recon → plan → research → draft → edit → seo → score, with optional hero image and derivatives stages.
  • GET /v1/jobs/:id exposes stage-by-stage status, timings, and metadata — no black box.
  • Deep-research mode: more research questions, stronger models on plan/draft/edit, 25 credits.
  • Strict quality rubric (clarity, hook, specificity, information gain); sub-80 articles are flagged in review.

Earlier work — auth, orgs, credit system — predates this log.