March 9, 2026

Review analytics dashboard, reviewRfp chat tool, expanded AI pipeline tracking, and E2E test coverage.

New admin tools for understanding AI review behavior, plus infrastructure improvements across testing and security.

New

  • Review analytics dashboard. A new admin page surfaces aggregate data from the dual-model RFP review step. See convergence rates, iteration distribution (1/2/3 attempts), per-category score comparisons between models, which model scores harsher, and expandable per-review detail with full iteration history. Supports filtering by organization and date range.

  • reviewRfp chat tool. Strutter AI can now review an RFP's quality on demand through the conversation. Ask "review this RFP" and it runs the dual-model review pipeline, returning scores across 8 categories with strengths, weaknesses, and improvement suggestions.

  • Branded 404 pages. Missing pages now show a custom 404 with a dizzy turkey illustration instead of a generic error screen.

  • Password reset. Users can now reset their password from the sign-in page. Clicking "Forgot password?" sends a branded reset email with a secure one-time link that expires in one hour.

Improved

  • RFP question generation accuracy. The AI wizard now generates the number of evaluation questions recommended by industry research (typically 75-125) instead of stopping at 20-30. The generation agent tracks progress toward the target count in real time and keeps generating until the goal is reached.

  • Expanded AI pipeline dashboard. The admin AI dashboard now tracks all AI job types (generation, scoring, review, enrichment) with org and date filtering, not just generation jobs.

  • API docs iframe support. The API documentation viewer now loads correctly when embedded in same-origin iframes.

  • reviewRfp strengths schema relaxed. The review tool no longer fails when the AI returns fewer strength items than the strict schema expected.

  • Social links in the footer. The site footer now includes links to Strutter's LinkedIn and X profiles, making it easier to follow product updates and announcements.

Fixed

  • RFP clarification jobs stuck in PENDING. A recent authentication change caused clarification jobs to remain in a pending state. The issue has been resolved, and jobs now process correctly.

  • RFP chat message ordering. Returning to an RFP after generation no longer places the RFP completion card above the conversation history. The clarification Q&A now appears in the correct order, including when opening the RFP on a different device or browser.

  • API documentation access for Pro plans. Pro tier organizations can now view the interactive API documentation tab on the API Keys page. Previously, only Enterprise organizations could access it despite Pro having full API access.

  • AI review now surfaces results. The dual-model quality review during RFP generation was silently failing, leaving the admin dashboard showing zeros and users with no feedback. Reviews now store failure metadata, show visible results (score, weak categories, and next-step options) in the conversation thread, and track failed reviews in the analytics dashboard.

  • "Review my RFP" routes correctly. Asking Strutter AI to review, evaluate, or assess your RFP now correctly triggers the quality review tool instead of misrouting to the content editor.

  • RFP wizard rate limiting. Creating or resuming an RFP no longer triggers "Too many requests" errors. Background status polling now uses a separate rate limit budget from AI endpoints and backs off gradually to avoid overloading the server.

  • Clarification history preserved during generation. The clarification Q&A exchange between the user and AI is now correctly passed through to the generation step. Previously, input validation silently dropped this context, which could result in the generated RFP not reflecting answers from the clarification conversation.

Security

  • Worker authentication hardened. The background task worker now requires authentication in all non-development environments with startup validation to ensure proper configuration.

  • API CORS origin control. The public API now supports a configurable origin allowlist. Defaults to open access for backward compatibility, but can be locked down to specific domains.

  • Rate limit monitoring. Rate limit blocks are now logged with structured data (IP, path, limiter type) for abuse pattern detection.

  • Clarification polling extended. The RFP wizard now waits up to ~20 minutes for the clarification step to complete, up from ~7 minutes. This prevents premature "Timed out waiting for clarification" errors when the backend research and AI calls take longer to process.

Under the Hood

  • Quality review co-reviewer upgraded to a more capable model for higher-fidelity consensus scoring.
  • E2E Playwright tests added for the full RFP generation lifecycle.
  • Iframe embedding support for API documentation routes.
  • Worker deploy timeout increased to prevent false failures during deployment.
  • Next.js bundle analyzer added to CI. Bundle size reports are uploaded as build artifacts for regression detection.
  • API docs viewer component simplified for improved security and reliability.
  • CI dependencies updated to latest stable versions.
  • Test mock type safety improved across the test suite. New typed factory builders cover every data model, and most unsafe type casts replaced with proper types.
  • Structured data updated from Product to SoftwareApplication schema across the marketing site, resolving Google Search Console warnings. Pricing page now includes SoftwareApplication and FAQPage structured data for richer search results.
March 9, 2026 | Strutter AI