Real deployed systems, real generated outputs, real infrastructure. This page documents the evidence that Q-Star AI is an operating AI-first company โ not a concept.
These systems are publicly accessible and running in production today.
Full-stack AI video production pipeline. Live and accessible. Generates short-form videos from prompts through script โ TTS โ render โ CDN delivery.
AI music composition live on WeChat. Users describe a mood or genre and receive a generated track. Accessible to WeChat's 1.3B users.
AI short video creation app. Submitted to App Store. Android APK builds available. Consumer-facing interface to Drama Engine infrastructure.
English-language company website with product pages, GCP use case, and evidence package. Hosted on production server with valid SSL.
Q-Star AI platform overview โ products, infrastructure, and Google Cloud migration plan.
Company introduction video covering all four products, the multi-agent infrastructure, and the Google Cloud migration plan. Generated using the Drama Engine pipeline.
All five reviewer-facing demo videos are rendered and publicly accessible. V4 AppForge and V5 GCP Migration completed 2026-05-13.
| Video | What It Proves | Google Cloud Relevance | Status |
|---|---|---|---|
| V1 โ Company Intro | Platform overview, 4 products, GCP migration plan | All GCP services overview | โ Public URL Ready |
| V2 โ Drama Engine Demo | Video pipeline: prompt โ script โ TTS โ render โ MP4 | Gemini, Cloud Run, GKE, Pub/Sub, Cloud Storage, GPU | โ Public URL Ready |
| V3 โ Multi-Agent Collaboration System | Codex-led orchestration with OpenClaw, Hermes, Claude, Q-Star agents, shared queues, R2/Qiniu handoff, and independent verification. Runtime evidence is on server 113. | Vertex AI, Cloud Run, Pub/Sub, BigQuery, Monitoring | โ Public URL Ready |
| V4 โ AppForge APK Generator โ | Prompt โ Kotlin code (950 lines) โ APK build โ 4 released APKs. QClip live on Android. Cloud Run + Gemini + Cloud Storage pipeline. | Gemini (code gen), Cloud Run (builds), Cloud Storage (APKs), BigQuery | โ Public URL Ready |
| V5 โ Google Cloud Migration โ | Current infra โ GCP 4-phase migration plan, service mapping, $3โ5K/mo usage projection, full evidence summary across 5 products. | Vertex AI, Cloud Run, Cloud Storage, BigQuery, Pub/Sub, Cloud Monitoring | โ Public URL Ready |
22+ completed video productions in R2 storage. Each production includes MP4 output, script, render plan, and summary โ proving the pipeline, not just the output file.
| Asset / Production | Files Present | What This Proves | GCP Relevance | Public Status |
|---|---|---|---|---|
| shengxia_fendela_ep1 commentary_en_v1/v2/v3.mp4 + summary.json |
MP4 (3 versions) ยท summary | Iterative video generation; multi-version output from same pipeline | Cloud Run, Cloud Storage, Gemini | Internal Evidence |
| yipinbuyi_ep1 commentary_en_v3.mp4 + summary.json |
MP4 ยท summary | English-language commentary generation from Chinese source | Cloud Run, Cloud Storage, Gemini | Internal Evidence |
| yipinbuyi_ep2 v4โv8 MP4 + srt + script + render_plan + highlights |
MP4 (5 versions) ยท SRT ยท script.md ยท render_plan.json ยท highlights | Full pipeline artifact set: script โ render plan โ SRT โ MP4; proves structured pipeline not ad-hoc | Gemini (script), Cloud Run (render), Pub/Sub (queue), Cloud Storage | Internal Evidence |
| ceo-darkest-secret_ep1 script + highlights + render_plan + summary + extraction |
script.md ยท render_plan.json ยท summary.json ยท video_content_extraction.json | Metadata-rich pipeline: content extraction โ script โ render plan โ structured output | Gemini (extraction + script), BigQuery (analytics) | Internal Evidence |
| ceo-darkest-secret_ep2 audio/s01โs05.wav + commentary_v1/v2.mp4 + metadata |
WAV audio (5 segments) ยท MP4 (2 versions) ยท metadata | TTS audio segments + assembled video; proves TTS โ render pipeline | Cloud Run (TTS), GKE (render), Cloud Storage | Internal Evidence |
| ceo-darkest-secret_ep3 commentary_v1.mp4 (23.5 MB) |
MP4 | Continued series production; pipeline handles multi-episode workflows | Cloud Storage, Cloud Run | Internal Evidence |
| tiktok_manor_ep1 commentary_v1.mp4 (48 MB) + summary.json |
MP4 (48 MB) ยท summary | Overseas-format video production; larger file size proves full-length render capability | Cloud Storage (large file), Cloud Run, GKE | Internal Evidence |
| apk-generator promo apk_promo_v1.mp4 (620 KB) |
MP4 | AppForge promotional video generated by Drama Engine; cross-product pipeline | Cloud Run, Cloud Storage | Needs Stable Public URL |
Total: 435+ MB generated video assets across 22+ files. Full inventory: r2-inventory.md
7 APK builds in R2 storage (136 MB total). Proves real app generation and packaging pipeline โ not mockups.
| File | Size | What This Proves | GCP Relevance | Public Status |
|---|---|---|---|---|
| QClip-v0.2.7-aigc.apk | 19.5 MB | Production APK build with AIGC features; real app packaging pipeline | Cloud Storage (artifact), Cloud Run (build worker) | Internal Evidence |
| QClip-v0.2.7-submit-aigc.apk | 20.2 MB | Store-submission variant; proves store packaging workflow | Cloud Storage, Cloud Run | Internal Evidence |
| QClip-v0.2.7-xiaomi-aigc.apk | 20.0 MB | Platform-specific build (Xiaomi store); proves multi-target build pipeline | Cloud Storage, Cloud Run | Internal Evidence |
| QClip-v0.3.0-360submit.apk | 20.0 MB | 360 store submission build; version progression proves active development | Cloud Storage, Cloud Run | Internal Evidence |
| QClip-v0.3.1-debug.apk | 20.3 MB | Debug build; proves CI/CD pipeline with debug/release variants | Cloud Storage, Cloud Run | Internal Evidence |
| qclip-debug-v0.1-aigc.apk | 15.9 MB | Early version; version history proves sustained development over time | Cloud Storage | Internal Evidence |
| qclip-v0.3.3-xiaomi.apk | 14.3 MB | Latest Xiaomi build; active version progression | Cloud Storage, Cloud Run | Internal Evidence |
Note: APK files are internal evidence. Direct public download not enabled. Available for reviewer access on request.
Structured pipeline artifacts proving Drama Engine is a real pipeline โ not just uploaded video files.
| File Type | Example | What This Proves | GCP Relevance |
|---|---|---|---|
| render_plan.json | yipinbuyi_ep2/render_plan.json | Structured render instructions generated by AI planner; proves LLM-driven pipeline orchestration | Vertex AI / Gemini (planning), Pub/Sub (dispatch) |
| script.md | ceo-darkest-secret_ep1/script.md | AI-generated narration script; proves LLM text generation is core to pipeline | Vertex AI / Gemini API |
| summary.json | shengxia_fendela_ep1/summary.json | Structured output metadata; proves pipeline produces machine-readable artifacts, not just media files | BigQuery (analytics ingestion) |
| video_content_extraction.json | ceo-darkest-secret_ep1/video_content_extraction.json | AI-extracted content from source video; proves multimodal understanding in pipeline | Vertex AI / Gemini (multimodal) |
| .srt subtitle file | yipinbuyi_ep2/*.srt | Auto-generated subtitles with timing; proves TTS โ subtitle sync pipeline | Cloud Run (TTS + subtitle worker) |
| highlights | yipinbuyi_ep2/highlights | AI-selected highlight segments; proves intelligent content curation layer | Vertex AI / Gemini (content analysis) |
Q-Star AI operates on self-managed cloud infrastructure today. This is the current state โ Google Cloud is the migration target.
Distributed across regions. Roles: gateway/networking (178), AgentOS brain (113), content processing (202), mobile backend (43.x).
What this proves: real multi-server infrastructure, not a single demo machineAll services containerized. Designed for GKE migration. Includes video pipeline, agent workers, API services, and mobile backend.
What this proves: container-ready workloads, GKE migration is straightforward435+ MB generated videos, 136+ MB APK artifacts, render plans, scripts, summaries. Active storage with ongoing writes.
What this proves: real generated output volume; Cloud Storage is the migration targetGemini is planned as the core Google model in the routing layer, with Claude, GPT, MiMo, and specialized models handled as external model routing where needed.
What this proves: the platform already has multi-model routing patterns, and Vertex AI / Gemini is the planned Google Cloud consolidation targetFull Google Cloud Startup Credits application package. Stored in R2; available for reviewer access.
| Document | Purpose | Status |
|---|---|---|
| google-cloud-credits-application.md | Application form answers โ positioning, workload profile, Vertex AI plan, usage projection | Ready |
| investor-brief.md | Investor-facing brief covering SAFE compliance, product matrix, GCP spend projection | Ready |
| qstar-architecture.md | Technical architecture: 4-server layout, model integrations, 7-phase GCP migration roadmap | Ready |
| qstar-usage-projection.md | 6-month and 12-month usage forecast with evidence basis | Ready |
| evidence-pack.md | Consolidated evidence summary for application reviewers | Ready |
| r2-inventory.md | Complete R2 asset inventory with public status and GCP relevance mapping | View โ |
Complete application materials, R2 asset access, and architecture documentation available on request.