.env.example

# Server Configuration
PORT=3000
NODE_ENV=development

# AI Service Configuration
# AI provider selection: "openrouter" (default) or "ollama"
AI_PROVIDER=openrouter

# OpenRouter Configuration (when AI_PROVIDER=openrouter)
OPENROUTER_API_KEY=your_openrouter_key_here
# Any OpenRouter text model - see https://openrouter.ai/models for options.
# Demo default matches gateway/CI/deployment docs; replace if needed.
OPENROUTER_MODEL=z-ai/glm-4.5-air:free
# Optional: override the OpenRouter endpoint (used in tests)
# OPENROUTER_URL=http://127.0.0.1:8080/api/v1/chat/completions

# Ollama Configuration (when AI_PROVIDER=ollama)
# OLLAMA_URL=http://localhost:11434
# OLLAMA_MODEL=llama2

# Payment Configuration
# Private key for the server wallet (recipient of payments) - REQUIRED
SERVER_WALLET_PRIVATE_KEY=your_private_key_here
# Recipient address embedded in payment contexts.
RECIPIENT_ADDRESS=0x2cAF48b4BA1C58721a85dFADa5aC01C2DFa62219
# Chain ID used by the gateway payment context (84532 = Base Sepolia)
CHAIN_ID=84532

# Token Configuration
# USDC_TOKEN_ADDRESS is reserved for future on-chain settlement and is not read by the current gateway/verifier.
# USDC_TOKEN_ADDRESS=
PAYMENT_AMOUNT=0.001

# Receipt Configuration
# Receipt storage backend: "redis" (default, survives restarts) or "memory" (tests/local only)
RECEIPT_STORE=redis
# Time-to-live for receipts in seconds (default: 86400 = 24 hours)
RECEIPT_TTL=86400

# Signature expiry configuration
# Chain ID enforced by the verifier (must match CHAIN_ID)
EXPECTED_CHAIN_ID=84532
# Reject legacy v1 authorization by default. Set to 1 only for an explicit rollback window.
MIN_AUTHORIZATION_VERSION=2

# Signature expiry window in seconds (default: 300 = 5 minutes)
SIGNATURE_EXPIRY_SECONDS=300

# Grace period for client clock skew in seconds (default: 60 seconds)
SIGNATURE_CLOCK_SKEW_SECONDS=60

# Verifier nonce replay store: "redis" for Docker/production (default), "memory" for local single-process
VERIFIER_NONCE_STORE=redis
VERIFIER_NONCE_KEY_PREFIX=microai:verifier:nonce:
VERIFIER_REDIS_TIMEOUT_MS=2000

# Service URLs (for Docker/production)
VERIFIER_URL=http://127.0.0.1:3002
# Required trusted public gateway origin used in request-bound authorization.
# Set only the origin (scheme + host + optional port), with no path/query/fragment.
PAYGATE_AUDIENCE=http://localhost:3000

# Maximum allowed request body size in bytes for verifier
# Default: 1048576 (1MB)
MAX_REQUEST_BODY_BYTES=1048576
# Comma-separated frontend origins allowed by the gateway CORS middleware.
# Use origins only, with no path/query/fragment.
# Defaults to http://localhost:3001 when unset or blank.
ALLOWED_ORIGINS=http://localhost:3001
# Comma-separated trusted proxy CIDRs/IPs for production deployments.
# Leave blank locally so Gin ignores X-Forwarded-For from untrusted clients.
TRUSTED_PROXIES=

# Rate Limiting
RATE_LIMIT_ENABLED=true

# Anonymous users (IP-based, no signature)
RATE_LIMIT_ANONYMOUS_BURST=5     # max burst tokens
RATE_LIMIT_ANONYMOUS_RPM=10      # requests per minute

# Standard users (signed requests)
RATE_LIMIT_STANDARD_BURST=20
RATE_LIMIT_STANDARD_RPM=60

# Verified users (future: premium tier)
RATE_LIMIT_VERIFIED_BURST=50
RATE_LIMIT_VERIFIED_RPM=120

# Cleanup interval for stale buckets (seconds)
RATE_LIMIT_CLEANUP_INTERVAL=300

# Request Timeout Configuration
# Global request timeout (seconds)
REQUEST_TIMEOUT_SECONDS=60
# AI endpoint timeout (seconds)
AI_REQUEST_TIMEOUT_SECONDS=30
# Verifier service timeout (seconds)
VERIFIER_TIMEOUT_SECONDS=2
# Health check timeout (seconds)
HEALTH_CHECK_TIMEOUT_SECONDS=2

# Metrics / Observability
# Enable the gateway Prometheus endpoint (default: true)
METRICS_ENABLED=true
# Gateway Prometheus endpoint path (default: /metrics)
METRICS_PATH=/metrics



# Redis Configuration (for Caching, Redis-backed Receipts, and verifier nonce storage)
# Use 'redis:6379' for docker-compose, 'localhost:6379' for local run
# REQUIRED when CACHE_ENABLED=true, RECEIPT_STORE=redis, or VERIFIER_NONCE_STORE=redis
REDIS_URL=redis:6379
REDIS_PASSWORD=
REDIS_DB=0

# Cache Settings
# Disabled by default so optional response cache writes cannot exhaust the
# receipt Redis instance. Enable only with dedicated cache capacity.
CACHE_ENABLED=false
# Time-to-live for cached items in seconds (default: 3600 = 1 hour)
CACHE_TTL_SECONDS=3600

# Frontend (Next.js / web/) — all read at build time, prefixed NEXT_PUBLIC_
# Verifier warm-up endpoint used by frontend banner
# Default: unset (set this to let the banner warm the verifier too)
# NEXT_PUBLIC_VERIFIER_URL=https://<verifier-app>.onrender.com
# Gateway URL the browser hits for /api/ai/summarize and /api/receipts/:id
# NEXT_PUBLIC_GATEWAY_URL=http://localhost:3000

# Chain ID the wallet widget expects (must match the gateway CHAIN_ID).
# Default 84532 (Base Sepolia).
# NEXT_PUBLIC_EXPECTED_CHAIN_ID=84532
# Display name for the expected chain — used by the wallet widget's
# "Switch to <name>" button + the summarize form placeholder copy.
# NEXT_PUBLIC_EXPECTED_CHAIN_NAME=Base Sepolia
# Pre-challenge fee label displayed under the summarize form. Informational
# only — the actual signed amount comes from the gateway's payment context.
# Default 0.001 / USDC.
# NEXT_PUBLIC_PAYMENT_AMOUNT=0.001
# NEXT_PUBLIC_PAYMENT_TOKEN=USDC
# PostHog Phase 1 browser analytics. Leave disabled locally unless you want
# event capture during manual testing.
# NEXT_PUBLIC_POSTHOG_ENABLED=false
# NEXT_PUBLIC_POSTHOG_PROJECT_TOKEN=phc_your_project_token
# NEXT_PUBLIC_POSTHOG_HOST=https://us.i.posthog.com