TechDemoForge v0.1.1

TechDemoForge v0.1.1 (2026-04-12)

This release hardens the data and API contract layers to make TechDemoForge more reliable as a local-first product and easier to maintain across backend/frontend boundaries.

Highlights

• SQLite-backed metadata persistence for projects, scenes, and jobs, while keeping binary assets on disk
• Job lifecycle recovery on restart for queued/running tasks
• Deterministic backend test suite coverage including end-to-end mock pipeline smoke tests
• OpenAPI export + TypeScript contract generation to reduce API drift risk
• CI guardrails that regenerate contracts and fail on drift

Persistence upgrades

• Projects and jobs are now persisted in metadata.sqlite3, improving queryability and recovery
• Scene-level status and asset path tracking is captured for more consistent lifecycle state
• Legacy JSON metadata files are still read for incremental migration, but no longer written
• Binary/media assets remain stored in backend/data/assets

Contract hardening

• OpenAPI schema is exported via backend/scripts/export_openapi.py
• Generated types live in shared/contracts/openapi.ts
• Frontend API client now uses the generated schema for response typing
• New command: cd frontend && npm run contracts:generate

Test & CI coverage

• New tests for metadata persistence, job recovery, and mock provider behavior
• End-to-end smoke test validates create → plan → generate → render in mock mode
• CI now regenerates OpenAPI types and fails if any drift is detected

Compatibility & notes

• No changes to provider isolation or mock mode behavior
• No changes to media storage paths or rendering pipeline behavior
• FFmpeg remains required for final MP4 output

---

TechDemoForge v0.1.0

TechDemoForge is a local-first tool for turning technical source material—READMEs, specs, architecture notes, product docs, and patent summaries—into polished short demo videos with storyboard editing, narration, scene planning, subtitles, and final export.

Highlights

• MiniMax-M2.7-backed text planning for storyboard generation
• Local-first mock mode for zero-key setup and deterministic development
• MiniMax provider paths for real text/speech/video workflows when configured
• Background job system for long-running generation and render tasks
• Docker-based quick start and persistent local project storage
• CI, tests baseline, README visuals, and contributor guidance

Current posture

• Mock mode remains the default runnable path
• Developers provide their own API keys
• Final MP4 rendering requires ffmpeg
• The project is optimized for technical honesty, local iteration, and extensible provider isolation rather than hosted SaaS complexity

Good fit for

• developer tools
• infrastructure products
• AI systems demos
• research or patent explainers
• technical product storytelling

Known limitations

• best suited today for local-first usage
• provider-backed behavior still depends on external API configuration and quotas
• early release; real-world workflow feedback will shape the next iteration

v0.1.0

TechDemoForge v0.1.0

TechDemoForge is a local-first tool for turning technical source material—READMEs, specs, architecture notes, product docs, and patent summaries—into polished short demo videos with storyboard editing, narration, scene planning, subtitles, and final export.

Highlights

• MiniMax-M2.7-backed text planning for storyboard generation
• Local-first mock mode for zero-key setup and deterministic development
• MiniMax provider paths for real text/speech/video workflows when configured
• Background job system for long-running generation and render tasks
• Docker-based quick start and persistent local project storage
• CI, tests baseline, README visuals, and contributor guidance

Current posture

• Mock mode remains the default runnable path
• Developers provide their own API keys
• Final MP4 rendering requires ffmpeg
• The project is optimized for technical honesty, local iteration, and extensible provider isolation rather than hosted SaaS complexity

Good fit for

• developer tools
• infrastructure products
• AI systems demos
• research or patent explainers
• technical product storytelling

Known limitations

• best suited today for local-first usage
• provider-backed behavior still depends on external API configuration and quotas
• early release; real-world workflow feedback will shape the next iteration