chore: add README, LICENSE (GPL 3.0), and update .env.example

- README.md: feature overview, tech stack, quick-start guide
- LICENSE: GNU General Public License v3.0
- .env.example: add ENCRYPTION_SECRET_KEY, LLM_RATE_LIMIT_RPM, correct model defaults

Authored by: Jack Levy

.env.example

@@ -10,6 +10,11 @@ PUBLIC_URL=
 # Generate: python -c "import secrets; print(secrets.token_hex(32))"
 JWT_SECRET_KEY=
 
+# Fernet key for encrypting sensitive user prefs (ntfy passwords, etc.)
+# Generate: python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"
+# Set once and never change after data has been written.
+ENCRYPTION_SECRET_KEY=
+
 # ─── PostgreSQL ───────────────────────────────────────────────────────────────
 POSTGRES_USER=congress
 POSTGRES_PASSWORD=congress
@@ -35,18 +40,23 @@ CONGRESS_POLL_INTERVAL_MINUTES=30
 LLM_PROVIDER=openai
 
 OPENAI_API_KEY=
-OPENAI_MODEL=gpt-4o
+OPENAI_MODEL=gpt-4o-mini
 
 ANTHROPIC_API_KEY=
-ANTHROPIC_MODEL=claude-opus-4-6
+ANTHROPIC_MODEL=claude-sonnet-4-6
 
 GEMINI_API_KEY=
-GEMINI_MODEL=gemini-1.5-pro
+GEMINI_MODEL=gemini-2.0-flash
 
 # For Ollama: use host.docker.internal to reach a locally running Ollama server
 OLLAMA_BASE_URL=http://host.docker.internal:11434
 OLLAMA_MODEL=llama3.1
 
+# Max LLM requests per minute (Celery enforces this globally across all workers).
+# Defaults: Gemini free=15, Anthropic paid=50, OpenAI paid=500.
+# Lower this if you hit rate-limit errors on a restricted tier.
+LLM_RATE_LIMIT_RPM=50
+
 # ─── Google Civic Information API ─────────────────────────────────────────────
 # Used for zip code → representative lookup in the Draft Letter panel.
 # Free tier: 25,000 req/day. Enable the API at: