# Q-Star AI Evidence Pack

## Positioning

Q-Star AI is building a multi-agent AI creation infrastructure for video, apps, and automated creative workflows.

Q-Star AI already operates multiple AI generation workflows across video, app generation, agent orchestration, and automated media production. We are preparing to consolidate these workloads on Google Cloud.

## Purpose

This evidence pack is designed for the Google Cloud Startup Credits application. It should prove that Q-Star AI is an operating AI platform with real workloads, deployed systems, demos, generated outputs, and cloud scaling needs.

## Live demo links

| Asset | URL | Evidence value |
| --- | --- | --- |
| Q-Star AI platform landing page | https://ai.q-star.ink/ | Public English company and platform positioning |
| Q-Star AI alternate domain | https://stu.q-star.ink/ | Existing deployed web endpoint |
| Demo asset domain | http://demo.q-star.ink/ | Hosted generated media and demo outputs |
| Drama platform domain | https://drama.q-star.ink/ | Existing AI content/product system to verify |

Before submission, verify each URL is reachable and capture screenshots.

## Existing deployed systems

| System | Evidence to collect |
| --- | --- |
| Q-Star AI website | Homepage screenshot, deployment path, Nginx config screenshot |
| QClip / AI video workflow | Prompt screen, generation logs, rendered video outputs |
| Drama Engine | Story pipeline screen, scene/task output, generated drama assets |
| APK Generator | App generation prompt, build logs, APK artifact screenshot |
| AgentOS | Agent planning screen, task graph, retry/review logs |
| OpenClaw Stack | Model routing logs, provider calls, queue dispatch evidence |

## Captured evidence snapshot

Captured on May 7, 2026 from the current 202 server.

| Evidence item | Verified state | Application value |
| --- | --- | --- |
| Public site | https://ai.q-star.ink/ returns HTTP 200 and serves the Q-Star AI landing page | Shows a deployed English platform presence |
| Evidence page | https://ai.q-star.ink/evidence.html and /evidence return HTTP 200 | Gives reviewers a direct evidence path |
| Nginx deployment | server_name ai.q-star.ink stu.q-star.ink; root /var/www/ai-qstar; HTTPS enabled by Certbot | Shows production-style web infrastructure |
| Running service | Nginx is active and enabled; configuration test passes | Supports reliability and operations claims |
| Deployed modules | agent.html, creative.html, drama-service.html, ecom-video.html, insight.html, shop.html | Shows multiple deployed product surfaces |
| Generated media outputs | Sample videos under /var/www/ai-qstar/videos for Douyin, Pinduoduo, Taobao, and Xiaohongshu | Supports AI video and generated output narrative |
| Docker workloads | huobao-drama, qstar-cobalt, yt-session-generator, kokoro-tts are running | Shows active backend and media-related workload infrastructure |
| Server capacity | 329 GB disk, 173 GB available, 15 GiB memory | Provides baseline capacity before Google Cloud migration |

## Server and log screenshots

Collect screenshots or exports for:

- Current server list and workload roles
- Running web/API services
- Agent workflow execution logs
- Model-routing logs
- Queue activity and async task counts
- Video/image/audio rendering worker logs
- APK build logs
- Storage usage screenshots
- API request volume
- Model token usage dashboard
- GPU worker utilization or job duration
- Error rate, health check, and uptime views

## Example generated outputs

Collect at least 5 to 10 examples:

- Generated short videos
- Short-drama scenes or trailers
- Generated images or visual assets
- Generated voice/audio/music assets
- Generated APK or app prototype
- Workflow audit trail or task graph
- Prompt-to-output demo recording

Each example should include:

- Input prompt
- Agent task plan
- Models or tools used
- Output file or URL
- Runtime or cost estimate if available

## Current API and workload evidence

The application should attach evidence supporting these baseline and projection numbers.

| Metric | Current baseline | 6-month projection | 12-month projection | Evidence source |
| --- | ---: | ---: | ---: | --- |
| Daily generations | 300 | 20,000 | 80,000 | API logs, generation database, workflow records |
| API requests / day | 5,000 | 350,000 | 1,500,000 | Gateway logs, server metrics |
| Async jobs / day | 1,000 | 120,000 | 500,000 | Queue logs, task database |
| Token consumption / month | 60M | 2B | 8B | Model provider dashboards |
| Video rendering minutes / month | 1,200 | 80,000 | 300,000 | Render worker logs |
| GPU inference hours / month | 120 | 3,000 | 10,000 | GPU server logs |
| Generated storage | 0.5 TB | 12 TB | 50 TB | Storage usage metrics |

These projections are based on current internal workloads, active demos, and planned early-access rollout.

## Google Cloud credits use case

We are applying for Google Cloud credits to support:

- Migration of agent workloads to Cloud Run and GKE
- Gemini API usage for text, code, planning, and multimodal generation
- Cloud Storage for generated videos, images, APKs, and prompt archives
- Pub/Sub for async task dispatch, rendering queues, retries, and workload fan-out
- BigQuery for usage analytics, cost reporting, and workload forecasting
- Cloud Logging and Cloud Monitoring for reliability, observability, and cost control
- GPU workloads for rendering, image and video processing, and batch generation

## Submission checklist

- Public English landing page
- Copy-ready Google Cloud application answers
- One-page pitch
- Cloud architecture diagram
- Usage projection
- Demo video script
- Evidence screenshots
- Current product/demo links
- 2-minute prompt-to-output demo video
- Real workload metrics exported from logs or dashboards
- Demo shot list: https://ai.q-star.ink/demo-shot-list.md

## Recommended final package order

1. Q-Star AI landing page
2. One-page pitch PDF
3. Google Cloud Startup Credits application answers
4. Cloud architecture diagram
5. Usage projection
6. Evidence pack
7. Demo video
8. Screenshot appendix