1 Commits

Author SHA1 Message Date
ozan fb6116b8b5 docs: meeting notes — Studio product vision (2026-05-22)
Capture platform strategy from live discussion: Gitea fork as bootstrap, single-domain model, services pipeline (git → GDD → credits → X Storage → agentic Godot pipeline), and temporary GitHub-like UI.

Co-Authored-By: Composer (Cursor) <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
2026-05-22 16:02:51 +01:00
@@ -0,0 +1,133 @@
# Meeting notes: Tinqs Studio product vision
**Date:** 2026-05-22
**Source:** Live discussion (mixed EN/TR transcript)
**Repo:** `tinqs-ltd/studio` (Gitea fork at `tinqs.com`)
---
## Purpose
Align on what **Tinqs Studio** is for, why we started from a **Gitea fork**, what is temporary UI vs long-term product, and how **services** (git, storage, asset generation, agentic pipeline) fit together under one domain.
---
## Platform model (GitHub analogy)
- **GitHub** puts marketing, login, org/repos, and logged-in product on **one origin** (`github.com`), not `dashboard.github.com`. Incognito vs logged-in: same URL, different surface.
- **Our platform** follows the same pattern: primary host is **`tinqs.com`** (not a separate `dashboard.tinqs.com`). Documentation can live on a docs host (e.g. `docs.tinqs.com`), similar to `docs.github.com`.
- **Advantage of owning our Git host:** organizations, secrets, auth, LFS, issues, and org docs already exist in-repo — we do not rebuild that stack from scratch.
---
## What Studio is (product definition)
- **Codebase to think in:** Studio is the **platform for agentic game development** — a specialized “harmonise” for making games with agents, not a generic dev portal.
- **Positioning (logged-out):** Pure HTML landing is intentional; copy can be rewritten freely. Example messaging discussed:
- *“Think Studio is the platform for agentic game development. Specialized AI agents work alongside your team, handling art, code, design, and production while you make the decisions that matter.”*
- “Build games with agents” (dollar-sign variants removed from copy).
- **Toolkit framing:** Motorik, Studio, bot, comms, and git hosting are **tools in one toolkit**, not unrelated products. Entry: **tools → toolkit → studio toolkit**.
- **Logged-out vs logged-in:** Marketing site when logged out; after login you should not see that page again — you land in the product (today that surface still looks like Gitea/GitHub).
---
## Why we started from Gitea (not greenfield)
Technical work we avoided by forking:
- Organization system, secrets, login/OAuth, large-file/LFS patterns
- **Git-native file truth:** assets and docs live in repos; Studio **views** them (HTML preview, explorer)
- **Keep long-term:** repo **inspector / file explorer** (browse repos, open files) — similar to needing an explorer in Claude; that shell stays
- **Replace over time:** everything else in the logged-in UI — current dashboard is a **bootstrap**, not the final IA
**Do not get stuck** on the current GitHub-like chrome; it is a starting point until content exists and we reshape groups into a real dashboard.
---
## Current state (honest)
- UI/UX is **wrong for end users** today; looks empty or “GitHub clone” until populated.
- **Design system** work in progress; favicons/logos generatable via Cursor from brand inputs.
- **CI/actions** still need wiring — active blocker mentioned for “making it work.”
- Developer-only lens so far; **designer/agent needs** can be added explicitly later.
---
## Services & monetization (discussed order)
| Layer | Role |
|-------|------|
| **Git / org hosting** | Base subscription; required before deeper game-asset workflows |
| **Story + git + game projects** | Services surfaced from main template (not final IA) |
| **GDD / design docs in repo** | Upload GDD (and related docs) to unlock **asset-generation** workflows |
| **Credits / stock** | Asset generation consumes credits; clean buy/sell model (e.g. tour conversion); can be invoked separately from full pipeline |
| **X Storage / Cloud Pro** | Long-term **cloud provider** for Godot-scale projects (100200 GB+ per game); few vendors want this; we become storage + server for games built on-platform |
| **Agentic pipeline (north star)** | Agent opens **local Godot project** (Unity/Unreal conversion out of scope); months-long iteration from GDD/CDD; 2D path deprioritized — Godot 3D + agentic from scratch |
**Godot focus:** Chosen because upstream already solved much of the git/LFS/hosting story; our job is **tour/port** and product layer, not re-proving git hosting.
---
## Domain & UX strategy
- **Prefer one brand domain** (subdomains OK for docs/API), not fragmented “cinema park” style splits — debated and rejected for now.
- **E-commerce analogy:** users often land on a **product/asset** page, not the dashboard; Studio navigation should reflect that over time.
- **Full-team mode:** version where whole teams can collaborate in one place (mentioned as direction).
---
## Developer & go-to-market angle
- Studio is the best place to **pull developers quickly** (familiar git surface, credits, chain/galvan-style demand mentioned).
- **SEO / AIO:** public repos and docs as discoverability surface (open-source + docs strategy).
---
## Integrations & migration (mentioned)
- Think Street **Go** code and transcripts (e.g. Mehmet) to land in this monorepo over time.
- Past pain: **~60 GB repo**, binary collisions — motivates LFS + X Storage story.
- Historical cloud spend: Unity Cloud, AWS, Google — informs “cheaper + Godot-native” positioning.
---
## Decisions (implicit)
1. **Gitea fork remains the foundation** for git, orgs, secrets, and file inspection.
2. **Logged-in Gitea UI is temporary**; marketing HTML and post-login dashboard will diverge strongly.
3. **Godot-only** for agentic pipeline; no commitment to Unity/Unreal import.
4. **Single-domain product** under `tinqs.com` family, not separate dashboard product URL.
5. **Git subscription before** heavy asset-generation / “asset monster” flows.
---
## Action items
| Owner | Task |
|-------|------|
| Platform | Finish **Gitea Actions/CI** wiring so deploy/build loop is trustworthy |
| Platform | Populate seed content so dashboard grouping can be designed from real objects |
| Design | Continue **design system** + replace favicon/logo set (brand generate pipeline) |
| Product | Rewrite logged-out **pure HTML** landing; stop treating Gitea home as final marketing |
| Product | Map **services** (git → GDD → credits → storage → agentic pipeline) into nav/IA |
| Team | Add **designer-facing agent requirements** alongside developer docs |
---
## Open questions
- Exact **credit pricing** and tour conversion mechanics
- Timeline to strip Gitea chrome down to **inspector-only** shell
- Where PM-style meeting notes should canonicalize (`pm` repo vs `studio/meetings/`)
---
## References
- [`ARCHITECTURE.md`](../ARCHITECTURE.md) — monorepo roles (`tinqs.com`, `docs.tinqs.com`, `bot.tinqs.com`)
- [`DESIGN.md`](../DESIGN.md) — brand asset pipeline
- Public docs: `docs.tinqs.com` (install, register, Cursor/DeepSeek guides)
---
*Transcript artifact at end of recording (“company born 1995”) — not treated as a product decision; likely ASR noise.*