Dragonflight - self-hosted broadcast media asset management. SRT/RTMP/SDI ingest via Blackmagic DeckLink, FFmpeg proxy generation, growing-file editing via SMB + Premiere Pro CEP panel, BullMQ job queue, S3-compatible storage (RustFS). Replaces Grass Valley AMPP FramelightX.
04ce096e67
Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3 |
||
|---|---|---|
| deploy | ||
| docs | ||
| services | ||
| .env.example | ||
| .gitignore | ||
| DESIGN.md | ||
| docker-compose.gpu.yml | ||
| docker-compose.worker.yml | ||
| docker-compose.yml | ||
| PRODUCT.md | ||
| README.md | ||
| setup-repo.sh | ||
Dragonflight
Self-hosted broadcast media-asset management system that replaces legacy tools like Grass Valley AMPP and FramelightX. Handles live ingest, growing-file editing, scheduling, transcoding, and asset management in a single operator-focused interface.
Repo renamed from
wild-dragon→dragonflight(2026-05-23). The old URL still redirects.
Home Dashboard
The home screen provides quick access to all major features and displays system status at a glance:
- Library — Browse projects, bins, and assets with hover-scrub previews
- Recorders — View configured capture devices and their status
- Editor — Timeline editor with cross-clip preview and render queue
- Jobs — Proxy and thumbnail queue with retry controls
- Settings — Configure storage, encoder, growing files, and capture SDK
- Dashboard — Operations view showing recent activity, job queue, and cluster health
Core Features
1. Live Ingest & Capture
Multi-protocol source capture with per-recorder codec settings
Dragonflight ingests from multiple sources simultaneously:
- SRT (Secure Reliable Transport) — caller and listener modes
- RTMP — standard streaming protocol
- SDI — via Blackmagic DeckLink cards with FFmpeg SDK 16.x patches
Each recorder can be configured with independent codec settings:
- ProRes (hi-res masters)
- H.264 / H.265 (proxies)
- DNxHR (Avid compatibility)
Audio routing and per-source configuration ensure flexibility for multi-camera productions.
2. Growing-File Editing
Live editing in Premiere Pro while capture is still writing
Editors mount the SMB landing zone directly in Premiere Pro and edit the live master file as it's being written. The included CEP (Custom Extension Panel) provides:
- Real-time clip detection and frame-accurate trimming
- One-click relink to final S3 master after promotion
- No waiting for capture to finish before editorial begins
3. Recorder Scheduler
Time-windowed recording automation
Schedule recordings with:
- One-shot, daily, or weekly recurrence
- Automatic start/stop via 15-second tick loop
- Conflict detection across recorders
- Project and bin assignment at schedule time
4. Library & Asset Management
Browse, search, and organize captured footage
The Library screen provides:
- Project and bin hierarchy
- Asset detail view with frame-anchored persistent comments
- Right-click context menu (move-to-bin, rename, delete)
- Global cmd/ctrl-K search across assets, projects, recorders, jobs, and users
- Hover-scrub preview with HLS playback
5. Jobs Queue
BullMQ-backed proxy and thumbnail generation
Automated background processing:
- Per-job retry logic with exponential backoff
- Bulk "retry all failed" for batch recovery
- Inline error messages with actionable diagnostics
- Status tracking: ingesting → processing → ready
Proxy encoder options:
- CPU-based: libx264 (H.264)
- GPU-accelerated: NVENC (NVIDIA) or VAAPI (AMD/Intel)
6. Timeline Conform & Export
FCP XML export with server-side FFmpeg rendering
The Premiere Pro panel exports FCP XML with:
- Server-side conform via FFmpeg
- Multiple output formats: H.264, H.265, ProRes
- Resolution presets: Broadcast, Web, Archive
- Batch processing with job queue integration
7. Hi-Res Auto-Relink
One-click batch relink of proxy clips to frame-accurate server-trimmed masters
After editing on proxies:
- Select clips in Premiere
- Trigger relink from the CEP panel
- Server trims hi-res segments to exact in/out points
- Concurrent trim worker pool for speed
- 24-hour TTL with automatic cleanup
8. Settings & Configuration
Centralized control for storage, encoding, and capture
Configure:
- S3 Storage — endpoint, bucket, credentials (with env-var fallback)
- Proxy Encoder — CPU vs GPU, bitrate, resolution
- Growing Files — SMB path, retention, auto-promotion
- Capture SDK — Blackmagic, AJA, or Deltacast uploader selection
9. Cluster & Distributed Capture
Primary + worker topology with remote DeckLink nodes
- Primary node runs API, scheduler, and web UI
- Worker nodes handle proxy/thumbnail jobs
- Remote capture nodes run DeckLink cards off-host
- Heartbeat health monitoring
- Automatic failover and recovery
10. Admin & User Management
Role-based access, token auth, and cluster monitoring
- User creation and role assignment
- API token generation for integrations
- Container and cluster node status
- System health dashboard
Quick Start
# Clone (repo renamed; old URL still redirects)
git clone https://forge.wilddragon.net/zgaetano/dragonflight.git
cd dragonflight
# Configure
cp .env.example .env
# Edit .env — S3 credentials + SESSION_SECRET at minimum
# Launch
docker compose up -d
# Open
open http://localhost:47434
Architecture
SDI / SRT / RTMP ──► capture (FFmpeg)
├─ HLS preview tee ──► /live/<assetId>/index.m3u8
└─ master output
├─ growing_enabled=true:
│ /growing/<projectId>/<clip>.mov
│ (Premiere mounts SMB, edits live)
│ └─► promotion worker uploads to S3
│
└─ growing_enabled=false:
multipart stream → S3
assets POST ──► proxy job ──► worker
├─ libx264 (CPU) or NVENC/VAAPI (GPU)
├─ thumbnail job
└─ status: ingesting → processing → ready
Tech Stack
- Runtime: Node.js 22, Docker Compose
- Backend: Express, PostgreSQL 16, Redis 7 + BullMQ
- Frontend: Vanilla React via in-browser Babel (no bundler), hls.js
- Media: FFmpeg 7.1 with SDK 16 DeckLink patches
- Codecs: ProRes, H.264, H.265, DNxHR, MOV/MP4/MXF containers
- Storage: S3-compatible (RustFS) for masters, proxies, thumbnails
Services
| Service | Port | Purpose |
|---|---|---|
| web-ui | 47434 | Browser SPA + capture controls |
| mam-api | 47432 | REST API + recorder orchestration + scheduler |
| capture | 47433 / 9000 / 1935 | DeckLink/SRT/RTMP ingest sidecar |
| worker | — | BullMQ proxy + thumbnail workers |
| db | 5432 | PostgreSQL 16 |
| queue | 6379 | Redis 7 |
Workflow Example: Live-to-Edit
- Operator schedules a recording on Recorder A for 14:00–15:30, assigns to "News/Segment-A" project
- Capture starts at 14:00, writes ProRes master to SMB landing zone
- Editor mounts SMB in Premiere, opens the live .mov file via the CEP panel
- Editor trims and marks in/out points while capture is still writing
- Capture finishes at 15:30, promotion worker uploads master to S3
- Editor clicks "Relink to Master" in CEP panel
- Server trims hi-res segment to exact in/out, stores for 24 hours
- Premiere relinks proxy clips to trimmed master
- Editor exports final timeline via FCP XML conform
Total time from end of capture to relinked master: ~2 minutes.
Operations
deploy/api-smoke.sh— verify every API endpoint after deploydeploy/onboard-node.sh— provision a remote worker hostdeploy/test-cluster.sh— primary↔worker connectivity smoke testdocs/GROWING_FILES_QUICKSTART.md— Premiere CEP panel install + growing-file flow
License
Proprietary — Wild Dragon LLC, all rights reserved.