scaffold: OpenScribe open-source self-hosted AI voice recorder

Bootstrap of the project (M0). Sets up the monorepo, design docs, hardware BOM,
the open API contract, component skeletons, licensing and CI, following the
Default Workflow SOP.

What changed:
- CLAUDE.md + docs/: copied the Default Workflow so sessions load the SOP.
- state/: PROJECT, ARCHITECTURE, DECISIONS, TODO, NOTES filled in for OpenScribe.
  ARCHITECTURE captures the four-part design (firmware, server, app, case) and the
  three sync paths; DECISIONS records the hardware, AI-stack, storage, app and
  licensing choices; TODO lays out milestones M1-M9.
- hardware/BOM.md: two build options (compact XIAO ESP32-S3 Sense; dev ESP32-S3 +
  I2S mic + SD), wiring/pinout, indicative cost.
- api/openapi.yaml: the completely open API (device + server surfaces), including
  recording list/download/delete and exports (wav/ogg/txt/srt/vtt/md/json).
- firmware/: PlatformIO ESP32-S3 project, two board profiles, pin map, boot scaffold
  with module seams for M1-M4.
- server/: FastAPI skeleton mirroring the OpenAPI, config for self-hosted MinIO,
  faster-whisper and Ollama; stub routes browsable at /docs.
- app/, case/: Flutter app plan; parametric OpenSCAD enclosure.
- Licensing: GPL-3.0 (code), CERN-OHL-S-2.0 (hardware), CC-BY-SA-4.0 (case/docs),
  REUSE-style LICENSES/ with SPDX headers; LICENSING.md explains the split.
- CI: Forgejo Actions workflow builds firmware (both profiles) and lints/imports server.

Why:
- Everything self-hosted and openly licensed per the user's requirements: an open
  API, three sync paths (BLE control, WiFi transfer, independent WiFi upload on
  charge to generic cloud storage), and a full self-hosted transcription+summary stack.

Notes:
- No custom PCB in v1; off-the-shelf modules. Physical verification waits on parts.
- Component code is stubs at M0; features land milestone by milestone, each as its
  own branch/PR per the workflow.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

server/README.md

@@ -0,0 +1,48 @@
+# OpenScribe server
+
+Self-hosted FastAPI server: ingests recordings, transcribes them (faster-whisper),
+summarises them (Ollama), and serves the open API with exports. Everything runs on
+hardware you own.
+
+> Status: M0 scaffold. The API shape is live and browsable at `/docs` with in-memory
+> stubs. Transcription lands in M5, summaries in M6, real storage/DB alongside.
+
+## Run (dev)
+
+```bash
+cd server
+python -m venv .venv
+. .venv/Scripts/activate     # Windows;  or:  . .venv/bin/activate
+pip install -r requirements.txt
+cp .env.example .env         # edit as needed
+uvicorn app.main:app --reload
+```
+
+- API docs (Swagger UI): http://localhost:8000/docs
+- Health: http://localhost:8000/health
+
+## Self-hosted dependencies
+
+For the AI features (M5/M6) you run, on your own kit:
+
+- **MinIO** (or WebDAV / NAS) for object storage - `OPENSCRIBE_STORAGE_BACKEND=s3`.
+- **Ollama** for summaries - `ollama serve` and `ollama pull llama3.1`.
+- **faster-whisper** downloads its model on first use; CPU works, CUDA is faster.
+
+None of these are required for plain recording and transfer; they add transcription and
+summaries.
+
+## Layout
+
+```
+app/main.py     FastAPI app + routes (mirrors ../api/openapi.yaml)
+app/config.py   Settings from env / .env
+app/models.py   Pydantic models (kept in sync with the OpenAPI schemas)
+requirements.txt
+.env.example
+```
+
+## API
+
+The contract is `../api/openapi.yaml`. The device implements the LAN "device" paths; this
+server implements ingest, transcript, summary and export.

server/app/__init__.py

@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: GPL-3.0-only
+"""OpenScribe self-hosted server package."""
+
+__version__ = "0.1.0"

server/app/config.py

@@ -0,0 +1,41 @@
+# SPDX-License-Identifier: GPL-3.0-only
+"""Server configuration, loaded from environment / .env (see .env.example).
+
+Everything points at self-hosted services by default: local object storage, a local
+Ollama, and a local faster-whisper model. Nothing here requires a proprietary cloud.
+"""
+from __future__ import annotations
+
+from pydantic_settings import BaseSettings, SettingsConfigDict
+
+
+class Settings(BaseSettings):
+    model_config = SettingsConfigDict(env_prefix="OPENSCRIBE_", env_file=".env")
+
+    # API
+    api_token: str = "change-me"          # bearer token for write endpoints
+
+    # Storage: "local" | "s3" | "webdav"
+    storage_backend: str = "local"
+    local_media_dir: str = "./media"
+
+    # S3-compatible (MinIO) - used when storage_backend == "s3"
+    s3_endpoint: str = "http://localhost:9000"
+    s3_bucket: str = "openscribe"
+    s3_access_key: str = ""
+    s3_secret_key: str = ""
+
+    # Metadata DB (SQLite to start; Postgres URL later)
+    database_url: str = "sqlite:///./openscribe.db"
+
+    # Transcription (faster-whisper)
+    whisper_model: str = "base"           # tiny|base|small|medium|large-v3
+    whisper_device: str = "cpu"           # cpu|cuda
+    whisper_compute_type: str = "int8"
+
+    # Summarisation (Ollama, self-hosted)
+    ollama_url: str = "http://localhost:11434"
+    ollama_model: str = "llama3.1"
+
+
+settings = Settings()

server/app/main.py

@@ -0,0 +1,68 @@
+# SPDX-License-Identifier: GPL-3.0-only
+"""OpenScribe server - FastAPI app.
+
+M0 scaffold: wires the routes from api/openapi.yaml with in-memory stubs so the API shape
+is real and browsable at /docs. The AI pipeline (faster-whisper transcription in M5,
+Ollama summaries in M6) and real storage/DB replace the stubs in later milestones.
+"""
+from __future__ import annotations
+
+from fastapi import FastAPI, HTTPException
+
+from .config import settings
+from .models import Recording, RecordingPage, Summary, Transcript
+
+app = FastAPI(
+    title="OpenScribe API",
+    version="0.1.0",
+    description="Self-hosted AI voice recorder server. See api/openapi.yaml.",
+)
+
+# In-memory store stands in for the DB + object storage until M5.
+_recordings: dict[str, Recording] = {}
+
+
+@app.get("/health")
+def health() -> dict:
+    return {
+        "status": "ok",
+        "storage_backend": settings.storage_backend,
+        "whisper_model": settings.whisper_model,
+        "ollama_model": settings.ollama_model,
+    }
+
+
+@app.get("/api/v1/recordings", response_model=RecordingPage, tags=["recordings"])
+def list_recordings(limit: int = 50) -> RecordingPage:
+    return RecordingPage(items=list(_recordings.values())[:limit], next_cursor=None)
+
+
+@app.get("/api/v1/recordings/{rec_id}", response_model=Recording, tags=["recordings"])
+def get_recording(rec_id: str) -> Recording:
+    rec = _recordings.get(rec_id)
+    if rec is None:
+        raise HTTPException(status_code=404, detail="No such recording")
+    return rec
+
+
+@app.get("/api/v1/recordings/{rec_id}/transcript", response_model=Transcript, tags=["server"])
+def get_transcript(rec_id: str) -> Transcript:
+    # Implemented in M5 (faster-whisper). Until then, signal "not transcribed yet".
+    if rec_id not in _recordings:
+        raise HTTPException(status_code=404, detail="No such recording")
+    raise HTTPException(status_code=409, detail="Not transcribed yet (M5)")
+
+
+@app.get("/api/v1/recordings/{rec_id}/summary", response_model=Summary, tags=["server"])
+def get_summary(rec_id: str) -> Summary:
+    # Implemented in M6 (Ollama). Until then, signal "not summarised yet".
+    if rec_id not in _recordings:
+        raise HTTPException(status_code=404, detail="No such recording")
+    raise HTTPException(status_code=409, detail="Not summarised yet (M6)")
+
+
+@app.post("/api/v1/ingest", response_model=Recording, status_code=202, tags=["server"])
+def ingest(recording: Recording) -> Recording:
+    # M5 will store audio to the object store and queue transcription + summary.
+    _recordings[recording.id] = recording
+    return recording

server/app/models.py

@@ -0,0 +1,59 @@
+# SPDX-License-Identifier: GPL-3.0-only
+"""Pydantic models mirroring api/openapi.yaml. Keep the two in sync."""
+from __future__ import annotations
+
+from enum import Enum
+
+from pydantic import BaseModel
+
+
+class SyncState(str, Enum):
+    local = "local"
+    uploaded = "uploaded"
+    ingested = "ingested"
+    transcribed = "transcribed"
+    summarised = "summarised"
+
+
+class Recording(BaseModel):
+    id: str
+    device_id: str | None = None
+    started_at: str
+    duration_s: float
+    sample_rate: int
+    channels: int
+    codec: str
+    size_bytes: int | None = None
+    sha256: str | None = None
+    source: str = "device"
+    sync_state: SyncState = SyncState.ingested
+    transcript_ref: str | None = None
+    summary_ref: str | None = None
+
+
+class RecordingPage(BaseModel):
+    items: list[Recording]
+    next_cursor: str | None = None
+
+
+class TranscriptSegment(BaseModel):
+    start: float
+    end: float
+    text: str
+    speaker: str | None = None
+
+
+class Transcript(BaseModel):
+    recording_id: str
+    language: str
+    model: str
+    text: str
+    segments: list[TranscriptSegment]
+
+
+class Summary(BaseModel):
+    recording_id: str
+    model: str
+    overview: str
+    key_points: list[str]
+    action_items: list[str]

server/requirements.txt

@@ -0,0 +1,19 @@
+# OpenScribe server dependencies.
+# Core API
+fastapi>=0.111
+uvicorn[standard]>=0.30
+pydantic>=2.7
+pydantic-settings>=2.3
+python-multipart>=0.0.9
+
+# Storage clients (self-hosted targets)
+boto3>=1.34            # S3-compatible (MinIO)
+webdavclient3>=3.14    # WebDAV / NAS
+
+# AI pipeline (self-hosted). Installed when M5/M6 land; listed here for reference.
+faster-whisper>=1.0    # transcription (CTranslate2)
+httpx>=0.27            # talk to Ollama for summaries
+
+# Dev
+pytest>=8
+ruff>=0.4