AI

BizDevBuddy Practice Zone

AI Practice Zone

ChangelogEnvironment: unknown

Changelog

The most important changes to the practice zone, newest first.

28 October 2026

One new AI engine available

  • GPT-6.1 Terra (Advanced tier, 20 credits) added to the regular prompt catalog.

14 October 2026

Three new AI engines available

  • Gemini 4 Argon (Expert tier, 30 credits), Muse Spark 1.3 (Advanced tier, 20 credits) and Muse Wonder (Expert tier, 30 credits) added to the regular prompt catalog.

6 October 2026

Security update

  • Updated the app's underlying framework packages to patched versions, resolving a known vulnerability that blocked deployments.

6 October 2026

Three new AI engines available

  • Claude 5.5 Sonnet (Advanced tier, 20 credits), GPT-6.1 Sol (Expert tier, 30 credits) and GLM-5.3 (Standard tier, 3 credits) added to the regular prompt catalog.

28 September 2026

Five new AI engines available

  • Claude 5.5 Opus (Expert tier, 30 credits), GPT-6 Sol and Luna, DeepSeek V4.1 Flash and Grok 4.7 added to the regular prompt catalog.

23 September 2026

Two new AI engines available

  • Claude Fable 5 and Claude Fable 5.1 (Expert tier, 30 credits) added to the regular prompt catalog.

16 September 2026

New image generator added

  • GPT-Image 2.5 is now available as an image generator.

14 September 2026

Stuck runs are now closed automatically

  • Runs that stay 'running' for more than 40 minutes are now closed automatically and shown as unsuccessful, instead of spinning forever.
  • Stuck runs from earlier days (since 19 August) have been cleaned up.

9 September 2026

Dark-mode logo and favicon

  • The EDUardo logo now switches to a light-on-dark version in dark mode.
  • The browser tab icon has been updated to the new EDUardo brand mark.

8 September 2026

Dark mode

  • A sun/moon switch in the top banner turns dark mode on or off. By default it follows your device setting, and your choice is remembered.
  • Course providers can optionally pin their course to light or dark mode.

7 September 2026

Course branding: provider logo in the top banner

  • The top banner now shows the course provider's own logo — the EDUardo logo for the Ethiopia AI Literacy Course.
  • An optional accent colour per course tints the banner highlights.

4 September 2026

Local copywrite draft & backtranslation prompt available

  • A new prompt is now offered in the practice zone for local copywrite drafting and backtranslation workflows.
  • Gemini 3.8 Flash (Standard Tier, 3 credits) added

3 September 2026

Three new AI engines available

  • Gemini 3.7 Flash, GLM-5.3-Flash, (Standard tier, 3 credits), Grok 4.6 (Advanced tier, 20credits) added to the regular prompt catalog.
  • Removed personalisation from LinkedIn-related workflows.

1 September 2026

Domain launch and cleaner variable hints

  • The Ethiopian AI Practice Zone is now live at https://ethiopai.practicezone.edu-simulation.com/.
  • Variable hints in the Run Service modal no longer duplicate the 'e.g.' prefix, so placeholders like [Topic: e.g. marketing emails] read cleanly.

27 August 2026

Waiting indicators and the nightly activity endpoint

  • Animated hourglass waiting indicator shown on the service page run transcript, in the chat history for in-flight runs, and a spinner on the run buttons while a run is being started.
  • First phase of the nightly activity retrieval endpoint (GET /api/v1/activity): paginated per-learner practice activity for a course, in the format the course platform's connector expects. No API key yet — protection follows in the next phase.

26 August 2026

Mandatory engines, real progress and live credits

  • Mandatory AI engine per course / stage / service: admins can lock a stage service to a single engine, and the run flow then only asks for confirmation instead of offering the full engine catalog.
  • Course-driven banner naming: the header shows the course's whitelabel name (e.g. EDUardo Practice Zone) with the course name underneath.
  • Real "Done!" badges and stage progress bars, derived from your actual run history instead of mock data.
  • Authoritative credit balance: the remaining credits on the main page now come from the backend profile, with a local estimate as fallback.
  • AI answers are rendered as formatted text (headings, lists, tables, code blocks) on the service page and in the chat history.

Open issues & plans

Known issues and planned developments that have not been addressed yet.

API Endpoints security to improve with IP whitelist and API keys

The activity retrieval endpoint is currently unprotected; add IP allowlisting and API-key/Bearer authentication before exposing it to external platforms.

LLM engine mandations

Mandatory engines are enforced per course / stage / service, but the surrounding admin tooling and backend sync are still being completed.

Workflow de-personalisations

Workflow runs still carry personalised inputs; de-personalisation is pending.

Missing workflows and prompts (4)

Stage 3 Agent, Stage 3 Community Problem Scan, Two-Tool Chain Starter, and Capstone Section Assembler are not yet available.