Back Office · office.temerarii.xyz
The machine — built + run on ourselves

← back to GTM Strategy

The machine — built + run on ourselves

We didn't buy the stack — we built it, and we run it on ourselves before we sell it.

Creative automation: We automated the ENTIRE creative process, end to end — ideation → script → storyboard → scene → render (Remotion + R3F, the Signal Field, heavy MG) → VO (Piper) + music + captions → 3 native aspects → publish. 342 video masters/yr + the daily long-form + the non-video lanes, briefed once. Most agencies automate a step; we automated the pipeline.

Creative control (driven like code): The generative + 3D stack is driven PROGRAMMATICALLY from Claude Code, not hand-clicked: ComfyUI (local GPU, RTX 4070/WSL2, via MCP/HTTP — on-brand imagery · LoRA · upscale, palette-locked to the §08 recipe; API-only, operator runs the app) + Blender MCP (text/image → 3D molds of anything → GLB → R3F/the Lab) + TouchDesigner (realtime generative visuals). So the creative pipeline — type, motion, 3D, imagery — is controlled like code.

AppSurfaceWhat
SlotSlot — the marketing command center (the product)slotslot (repo)the productized platform: modular · agent-powered · MCP-first · 7 integrated sub-MVPs, research → publishing. The temerarii-studio engine (4-4-5 · output policy · cadence · per-asset gate · content-index) is its spec; this office is the same machine run on ourselves first.
The content machine (back office)office.temerarii.xyzgoal → stages → gates harness that plans, composes, gate-checks + publishes the whole calendar — the operator engine + the run-on-ourselves proof behind Slotted
SIE — Sovereign Intent Engineapp.siedata.devthe consent/data-rights wedge product — compliance-as-a-service + declared-intent capture
The Labtemerarii.xyzreal-time WebGL proof-of-capability across the 8 pillars
Custom enrichment stackengine/our own Clay-style scrape→enrich→score outbound pipeline, built from MCPs we control
The content-creation process: Plan → script → scene → render → sound → host → publish, every step gate-checked. 342 video assets/year from one planning pass.
  1. PLAN — topics + calendar + GTM → content-index (one source, 342 assets: 94 cells · 30 long-form · 210 threads · 8 holidays)
  2. SCRIPT — beat arc (open·legacy·teach·proof·futurist·resolve) authored in the locked voice; caption = source → VO + on-screen card derive
  3. SCENE — scene-paced (≥5s/scene, minimums per tier), heavy MG every scene, correlated script→scene→transition→music→VO
  4. RENDER — Remotion: SceneReel (social) + LongFormChapters (5–10min YouTube), 3 native aspects, gl:angle
  5. SOUND — Piper VO (free/local) + ducked music bed + burned captions, −16 LUFS
  6. HOST + PUBLISH — Cloudinary → /media review → Blotato 9-channel + email/blog/SMS (staging until GO)

Rules enforced — verify_scene_pacing · verify_rendered_duration (minimums) · the 10× readiness gauntlet (0 errors) · verify_variety (no redundancy) · the brand rails

Masters/yr 342Long-form >=1 YouTube long-form/week (30+ over W23-52), each spinning off ~5 shortsPer master 3 native aspects (9:16·1:1·16:9) + 4K + the non-video lanes (email·blog·SMS)Published W23-W39 ≈ 5,550 published touches from the masters

← back to GTM Strategy