Changelog

All notable changes to Agent Probe.

2026-11-02

New Domain & Single Sign-On with Firebase

v0.8.0
  • Agent Probe is now available at agentprobe.ai
  • Sign in with your Google account via Firebase Authentication, exchanged for a secure session token on our server
  • Website refresh: AntrekLabs company details, LinkedIn link and a contact form that opens your email app
  • Privacy policy updated with KVKK data controller information
2026-05-12

Not In Scope Status, Report Exports & Score Bucketing

v0.7.0
  • New Not In Scope result status — scope refusals are tracked separately from failures
  • Multi-select status filter with invert and clear actions
  • Score-based bucketing (fail / partial / pass) for result chips, badges and filters
  • Report polish: improved recommendations, complete CSV/JSON exports and a print-friendly PDF layout
  • Large dashboard responses are now gzip-compressed for faster loading
  • Support for large grounded test datasets generated from customer documents
2026-03-23

Tutorials Library, SEO & Language Routing

v0.6.0
  • Tutorials video library and demo videos in English and Turkish
  • URL-based language routing (/en/… and /tr/…)
  • SEO and AI-indexing improvements (structured data, sitemap, llms.txt) and a new favicon
  • Video playback and thumbnail fixes; large videos split for reliable streaming
2026-03-18

Custom Agents & Full Turkish Website

v0.5.0
  • Custom agents: define your own HTTP endpoint agents (request template, response field parsing, connection test) and select them in the model dropdown
  • Website and dashboard served from a single server (/ website, /app dashboard) with clean URLs and auth-aware navigation
  • Complete Turkish translation of the documentation and website
  • Language-aware demo video and improved mobile menu
2025-03-15

Guided Tour, User Management & Chat Filters

v0.4.0
  • Added interactive guided tour for new users
  • User management modal for admin users (approve / reject signups)
  • Chat panel filter improvements — search and filter test results
  • Improved dashboard responsive layout
2025-03-10

Profile Menu, API Keys & Webhook Improvements

v0.3.0
  • Profile menu with settings, API key management, and password change
  • API key generation (ap-xxxx... format) for programmatic access
  • Signup approval workflow — new users require admin approval
  • Webhook panel improvements with cron scheduling
  • JWT-based authentication system
2025-02-25

Test Comparison & History

v0.2.0
  • Test history panel — view and reload past test runs
  • Side-by-side test comparison (select 2 tests to compare)
  • Evaluator-level diff analysis between test runs
  • Improved real-time streaming of test results
  • Configuration strip showing active test parameters
  • Turkish language support for dashboard UI
2025-02-10

Initial Release

v0.1.0
  • 6-layer test pyramid with 16 evaluators
  • Golden Datasets in English and Turkish (MMLU, TruthfulQA, BBQ, ToxiGen, JailbreakBench, PAWS, AdvGLUE, NeedleBench)
  • Split-screen dashboard: chat + real-time evaluation
  • 300+ LLM models via OpenRouter integration
  • Manual and automated test modes
  • Docker deployment with Google Cloud Run support
  • JSON and HTML report generation