TestHub Planet Edition: Full Release History and Evolution Insights

Initial release – platform refactor

Technology stack upgrade : Django upgraded from 4.2.7 to 6.0.2; Celery replaced by lightweight Django‑Q2. requirements.txt reduced from 192 lines to 49, cutting three‑quarters of dependencies.

Deployment simplification : a single start.sh script launches backend, frontend, workers and WebSocket without manual virtual‑environment activation.

Performance monitoring : Admin panel adds request‑performance logging with colour‑coded response‑time bands (red = slow, orange = medium, green = normal) and nightly aggregation.

Notification templates : fully configurable, supporting custom content, email and variable substitution.

V1.1 – AI integration and module unification

Knowledge‑base test‑case generation : supports mixed, semantic and keyword search; AI model generates test cases that can be saved directly or exported to Excel.

Axure link parsing : entering an online Axure link triggers asynchronous conversion of prototype pages into structured Markdown requirements, which are then used to generate test cases. Closing the browser does not affect task progress.

Excel batch import : uses the Django‑Q asynchronous framework; import progress is polled in real time and errors are reported with precise row numbers and reasons.

Unified project management : a single entry in the configuration centre links UI automation, API testing and APP testing modules, eliminating data silos.

V2.0 – UI overhaul, engine upgrade and reporting

UI refresh : left‑side menu changed from dark to transparent; dark‑mode theme added for night‑time work; configuration centre includes an orange theme switch.

Document parsing : supports over 1,000 formats (PDF, Word, TXT, Markdown, XMind, ZIP archives, etc.) with automatic recursive handling of compressed packages.

OCR integration : default Tika service and optional custom large‑model option for image‑to‑text extraction.

Support for Swagger, Postman and HAR batch import.

Variable extraction via JSONPath and regular expressions.

Response bodies displayed as a tree with JSONPath query capability.

Test‑suite conditions using Python expressions; independent variable/assertion configuration.

Report engine upgraded to Allure‑Pytest, supporting online view and offline download.

UI/APP automation : all modules integrated with Allure reports; APP module adds remote real‑time screen control, laying groundwork for cloud‑device testing.

V3.0 – Scale, operations tooling and AI assistant

APP automation enhancements : remote control, device‑performance monitoring, application‑performance monitoring and debug‑log viewing.

Startup mechanism : switched to Uvicorn, allowing HTTP, SSE and WebSocket to share a single port, removing the need for multiple terminal windows.

Operations toolbox : central management of environment categories, SSH credentials and middleware connections; real‑time log querying with keyword search for MySQL, Redis and MongoDB.

Global AI assistant : floating button in the lower‑right corner provides platform Q&A roadmap targets natural‑language‑driven test execution.

Knowledge‑base engine refactor : migrated to Tika Server, removing unstable dependencies (LibreOffice, PyMuPDF); parsing speed increased, token consumption reduced; supports multi‑file uploads and incremental additions.

Interface testing refinements : one‑click copy/delete, drag‑drop ordering, batch import/export optimizations, suite case sorting.

V4.0 – Hermes digital human

Hermes is a native “quality digital human” that executes a full test‑lifecycle workflow from a single natural‑language command: write test cases, run tests, build plans, assemble suites, query data and generate test data.

Hermes bundles 17 tightly integrated tools covering:

Project information query.

Test‑case creation and query with multi‑dimensional filtering.

Test‑plan management (view plans, execution statistics).

Full‑chain API testing (request creation → suite assembly → asynchronous execution).

Test execution and reporting (run → update → report → failure analysis).

Data factory for bulk test‑data generation.

Defect tracking (failed‑case lookup, result analysis, defect linking).

The workflow is a seamless chain: project → case → supplement → execute → update → report → analyze.

Hermes embeds Playwright MCP, exposing 27 browser‑automation capabilities, enabling “describe it, it executes” behaviour.

Hermes operates as a native part of the platform, with read/write access to platform data rather than as an add‑on.

Consistent principles

Visible update frequency

From the initial version to V4.0, releases have occurred regularly, each introducing substantial new functionality rather than superficial version bumps.

User‑driven development

Features such as Excel import, ops tools, interface drag‑drop ordering and dark mode were added directly after community requests were evaluated and approved.

AI‑first strategy

Each version leveraged AI to automate repetitive testing tasks—knowledge‑base semantic search, OCR with large models, document parsing, global AI assistant, and finally the Hermes digital human—realizing tangible efficiency gains.