# =============================================================================
# InsForge Environment Configuration
# =============================================================================
# Copy this file to .env and fill in your values:
#   cp .env.example .env
#
# Security Notes:
# - Never commit .env to version control
# - Use strong, unique secrets in production
# - Rotate secrets regularly
# =============================================================================

# Compose project name — prefixes every container, volume and network, so keep
# it stable across upgrades. A second instance on the same host needs its own
# value here, or `docker compose up` will adopt this one's containers.
COMPOSE_PROJECT_NAME=insforge

# Which compose file `docker compose` runs. Below is the development stack; it
# has to stay equal to what Compose would discover on its own in this directory. A self-hosted install points it at the image-only compose
# file instead — deploy/setup.sh writes that for you:
#   COMPOSE_FILE=deploy/docker-compose/docker-compose.yml
# Add a storage overlay by appending it with a colon:
#   COMPOSE_FILE=deploy/docker-compose/docker-compose.yml:docker-compose.minio.yml
COMPOSE_FILE=docker-compose.yml

# -----------------------------------------------------------------------------
# Server Configuration
# -----------------------------------------------------------------------------
PORT=7130

# Max body size for JSON payloads (default: 100mb)
# High default ensures "out-of-the-box" reliability for large metadata/storage requests.
# Users can decrease this for hardened security on low-resource environments.
MAX_JSON_BODY_SIZE=100mb

# Max body size for URL-encoded payloads (default: 10mb)
MAX_URLENCODED_BODY_SIZE=10mb

# Keep-alive timeout for idle HTTP connections in milliseconds (default: 65000)
# Must exceed the idle timeout of any load balancer or proxy in front of the
# backend so clients never reuse a connection the server has already closed.
KEEP_ALIVE_TIMEOUT_MS=65000

# Max concurrent connections from the backend to PostgREST (default: 50)
# Tune per instance size, and keep it aligned with PostgREST's database pool
# (PGRST_DB_POOL): raising it beyond the pool only moves queuing into PostgREST.
POSTGREST_MAX_SOCKETS=50

# Idle keep-alive connections retained toward PostgREST (default: 10)
POSTGREST_MAX_FREE_SOCKETS=10

# How long an idle keep-alive connection toward PostgREST is kept before the
# backend closes it (default: 4000). Must stay below PostgREST's server-side
# idle timeout so the backend always closes idle sockets first — reusing a
# socket the server already closed causes ECONNRESET ("socket hang up").
POSTGREST_FREE_SOCKET_TIMEOUT_MS=4000

# PostgREST's own database connection pool. Keep aligned with POSTGREST_MAX_SOCKETS.
# PGRST_DB_POOL=50

# Public URL PostgREST advertises in its OpenAPI output.
# POSTGREST_OPENAPI_SERVER_PROXY_URI=

# -----------------------------------------------------------------------------
# PostgreSQL Configuration
# -----------------------------------------------------------------------------
# These are optional - defaults are shown below
# In production, use strong passwords and consider external database services
POSTGRES_USER=postgres
POSTGRES_PASSWORD=postgres
POSTGRES_DB=insforge

# -----------------------------------------------------------------------------
# Ports (Configurable)
# -----------------------------------------------------------------------------

# PostgreSQL
POSTGRES_PORT=5432

# PostgREST API
POSTGREST_PORT=5430

# Main application ports
APP_PORT=7130
AUTH_PORT=7131
UI_PORT=7132

# Deno serverless runtime
DENO_PORT=7133

# API Base URLs - Update if running on different host/port
API_BASE_URL=http://localhost:7130
VITE_API_BASE_URL=http://localhost:7130

# -----------------------------------------------------------------------------
# Authentication & Security
# -----------------------------------------------------------------------------
# JWT_SECRET: Must be at least 32 characters. Use a secure random generator!
# Example: openssl rand -base64 32
JWT_SECRET=your-secret-key-here-must-be-32-char-or-above

# Root admin credentials - CHANGE THESE IN PRODUCTION!
ROOT_ADMIN_USERNAME=admin
ROOT_ADMIN_PASSWORD=change-this-password

# Encryption key for secrets and database encryption
# IMPORTANT: Set this to a separate 32+ character secret from JWT_SECRET.
# If not set, JWT_SECRET is used as fallback — but rotating JWT_SECRET will
# permanently corrupt all stored secrets (API keys, OAuth tokens, etc.).
# Generate with: openssl rand -base64 32
ENCRYPTION_KEY=

# API key for authenticated requests ('ik_' prefix), and anon key for public
# clients ('anon_'). Leave empty and the backend generates them on first boot;
# deploy/setup.sh generates them for you. Whatever is here is used as-is, so do
# not leave a placeholder — it becomes this instance's real key.
ACCESS_API_KEY=
ACCESS_ANON_KEY=

# Cloud API Host (Optional)
# Only needed if using cloud features
CLOUD_API_HOST=https://api.insforge.dev

# -----------------------------------------------------------------------------
# Deployment Configuration (Optional)
# -----------------------------------------------------------------------------
# Required for self-hosted site deployments and custom domains.
# Legacy deployments also require AWS_S3_BUCKET to be configured.
VERCEL_TOKEN=
VERCEL_TEAM_ID=
VERCEL_PROJECT_ID=

# -----------------------------------------------------------------------------
# AWS Config Bucket Configuration (Optional)
# -----------------------------------------------------------------------------
# Used for loading remote configuration files from S3
# If not provided, defaults will be used
AWS_CONFIG_BUCKET=insforge-config
AWS_CONFIG_REGION=us-east-2



# -----------------------------------------------------------------------------
# Storage Configuration
# -----------------------------------------------------------------------------
# Storage options:
# - Local filesystem: leave S3_BUCKET empty (default)
# - Any S3-compatible store (AWS S3, MinIO, RustFS, Wasabi, R2, Tencent COS,
#   Aliyun OSS ...): set the S3_* variables. For AWS S3, leave S3_ENDPOINT_URL
#   empty (SDK default endpoints; credentials may also come from an IAM role).
#
# Local storage directory (default: ./insforge-storage relative to cwd)
# In Docker this is set to /insforge-storage automatically via the Dockerfile.
# PaaS deployments (Zeabur, Render, etc.) should set this to match the
# persistent volume mount path.
# STORAGE_DIR=./insforge-storage
#
# Bucket name — setting this selects the S3 storage provider
S3_BUCKET=
# SigV4 signing region (default us-east-2; any value works for MinIO/RustFS —
# gateway clients must sign with the same region)
S3_REGION=
# Store credentials
S3_ACCESS_KEY_ID=
S3_SECRET_ACCESS_KEY=
# Custom endpoint for non-AWS stores (leave empty for AWS S3)
S3_ENDPOINT_URL=

# Max upload file size in bytes (default: 52428800 = 50MB)
# Examples: 10485760 = 10MB, 104857600 = 100MB
MAX_FILE_SIZE=
# Path-style addressing (default true, required by MinIO). Set to false for
# providers that require virtual-hosted-style (Tencent COS, Aliyun OSS).
S3_FORCE_PATH_STYLE=true
# Presigned URLs (default true): clients upload/download directly against the
# S3 endpoint. Set to false for PROXY MODE: all object bytes stream through
# the backend instead. Required when the endpoint is not reachable by
# browsers (bundled MinIO/RustFS on the Docker network) or lacks POST-policy
# support (Cloudflare R2). In proxy mode uploads are capped by the storage
# max-file-size setting; larger objects go through the S3 gateway.
S3_USE_PRESIGNED_URLS=true
# Max single S3-gateway upload in bytes (default 5368709120 = 5GB)
S3_MAX_OBJECT_SIZE_BYTES=

# ⚠️ AWS_* variables are for InsForge Cloud projects only — cloud provisioning
# sets them on managed instances. For self-hosting, use the S3_* variables
# above and leave these untouched. (AWS_S3_BUCKET / AWS_REGION are the legacy
# names and still work as fallbacks for S3_BUCKET / S3_REGION;
# AWS_ACCESS_KEY_ID / AWS_SECRET_ACCESS_KEY are AWS-account credentials used
# by CloudWatch logging and CloudFront.)
AWS_ACCESS_KEY_ID=
AWS_SECRET_ACCESS_KEY=
AWS_REGION=
AWS_S3_BUCKET=

# Bundled object store (optional): run MinIO or RustFS next to InsForge by
# appending its overlay to COMPOSE_FILE above. Proxy mode is preconfigured, so
# no S3_* settings below are needed:
#   COMPOSE_FILE=deploy/docker-compose/docker-compose.yml:docker-compose.minio.yml
# Override the default store credentials (DO change these in production):
# MINIO_ROOT_USER=insforge
# MINIO_ROOT_PASSWORD=insforge-minio-secret
# RUSTFS_ACCESS_KEY=insforge
# RUSTFS_SECRET_KEY=insforge-rustfs-secret
# Publish the MinIO admin console on the host (off by default):
# MINIO_CONSOLE_PORT=9001

# CloudFront signed URLs (InsForge Cloud projects only — self-hosting should
# leave these untouched)
# When AWS_CLOUDFRONT_URL is set, storage downloads are served through
# CloudFront with signed URLs instead of S3 presigned URLs. Ignored when
# S3_ENDPOINT_URL is set (CloudFront does not front S3-compatible providers).
# Setup: https://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/private-content-trusted-signers.html
#
# AWS_CLOUDFRONT_URL: CloudFront distribution domain (e.g. https://d1234abcd.cloudfront.net)
# AWS_CLOUDFRONT_KEY_PAIR_ID: Trusted key group key-pair ID (e.g. K2JCJMDEHXQW5F)
# AWS_CLOUDFRONT_PRIVATE_KEY: PEM private key. Use \n for newlines when kept on one line.
# If URL is set but key pair or private key is missing, downloads fall back to S3 presigned URLs.
AWS_CLOUDFRONT_URL=
AWS_CLOUDFRONT_KEY_PAIR_ID=
AWS_CLOUDFRONT_PRIVATE_KEY=

# -----------------------------------------------------------------------------
# Logging Configuration
# -----------------------------------------------------------------------------
# CloudWatch Logging: Set AWS_REGION + AWS credentials above
# File-based Logging: Configure LOGS_DIR below (used when AWS credentials not provided)
# Directory where log files will be stored (defaults to ./logs)
LOGS_DIR=

# -----------------------------------------------------------------------------
# Anonymous Telemetry
# -----------------------------------------------------------------------------
# InsForge sends anonymous self-host usage events so maintainers can understand
# active deployments, version adoption, and deployment success. It never sends
# secrets, environment variables, logs, domains, file paths, project names, or
# database contents.
#
# Disable all telemetry:
# INSFORGE_TELEMETRY_DISABLED=1


# -----------------------------------------------------------------------------
# AI/LLM Configuration
# -----------------------------------------------------------------------------
# OpenRouter compatibility credential. On first startup, the value is copied into
# InsForge's encrypted secret store and remains a fallback when no stored key exists.
# Model Gateway Settings can replace it; stored values always take precedence.
# Get an API key from https://openrouter.ai/keys
OPENROUTER_API_KEY=

# Maximum output tokens per chat completion request (default: 16384).
# Increase this for models with larger output windows (e.g. 32768 for 32K-output models).
# Must be a positive integer. Invalid values are silently ignored and the default is used.
MAX_COMPLETION_TOKENS=16384

# -----------------------------------------------------------------------------
# Stripe Payments Configuration (Optional)
# -----------------------------------------------------------------------------
# Developer-owned Stripe secret keys. Test key is used for implementation and
# validation. Live key is used only for explicit go-live flows.
STRIPE_LIVE_SECRET_KEY=
STRIPE_TEST_SECRET_KEY=

# -----------------------------------------------------------------------------
# Razorpay Payments Configuration (Optional)
# -----------------------------------------------------------------------------
# Developer-owned Razorpay keys. Test keys are used for implementation and
# validation. Live keys are used only for explicit go-live flows.
RAZORPAY_LIVE_KEY_ID=
RAZORPAY_LIVE_KEY_SECRET=
RAZORPAY_TEST_KEY_ID=
RAZORPAY_TEST_KEY_SECRET=

# -----------------------------------------------------------------------------
# Analytics Configuration
# -----------------------------------------------------------------------------
# PostHog - Only needed for local development
# Get your key from https://posthog.com/settings/project
VITE_PUBLIC_POSTHOG_KEY=

# -----------------------------------------------------------------------------
# OAuth Configuration (Optional)
# -----------------------------------------------------------------------------
# Enable social login by configuring one or more providers below.
# Each provider requires registering an application in their developer console.
#
# Google OAuth
# Console: https://console.cloud.google.com/
# Redirect URI: http://localhost:7130/auth/google/callback
GOOGLE_CLIENT_ID=
GOOGLE_CLIENT_SECRET=

# GitHub OAuth
# Console: https://github.com/settings/developers
# Redirect URI: http://localhost:7130/auth/github/callback
GITHUB_CLIENT_ID=
GITHUB_CLIENT_SECRET=

# Microsoft OAuth (Azure AD)
# Console: https://portal.azure.com/
# Redirect URI: http://localhost:7130/auth/microsoft/callback
MICROSOFT_CLIENT_ID=
MICROSOFT_CLIENT_SECRET=

# Discord OAuth
# Console: https://discord.com/developers/applications
# Redirect URI: http://localhost:7130/auth/discord/callback
DISCORD_CLIENT_ID=
DISCORD_CLIENT_SECRET=

# LinkedIn OAuth
# Console: https://www.linkedin.com/developers/apps
# Redirect URI: http://localhost:7130/auth/linkedin/callback
LINKEDIN_CLIENT_ID=
LINKEDIN_CLIENT_SECRET=

# X (Twitter) OAuth
# Console: https://developer.twitter.com/en/portal/dashboard
# Redirect URI: http://localhost:7130/auth/x/callback
X_CLIENT_ID=
X_CLIENT_SECRET=

# Apple OAuth (Sign in with Apple)
# Console: https://developer.apple.com/account/resources/identifiers/list
# APPLE_CLIENT_ID: Your Services ID (e.g., com.yourapp.service)
# APPLE_CLIENT_SECRET: JSON string with teamId, keyId, and privateKey
#   Format: {"teamId":"XXX","keyId":"YYY","privateKey":"-----BEGIN PRIVATE KEY-----\n..."}
# Redirect URI: http://localhost:7130/auth/apple/callback
APPLE_CLIENT_ID=
APPLE_CLIENT_SECRET=

# -----------------------------------------------------------------------------
# Multi-tenant Cloud Configuration
# -----------------------------------------------------------------------------
# These are only used for the cloud-hosted solution
# Leave empty for self-hosted deployments.
#
# PROJECT_ID is safe to set when self-hosting: it scopes compute services and edge
# function deployments, and is not taken as a signal that this is a cloud-managed
# project. What marks a deployment as cloud-managed is AWS_INSTANCE_PROFILE_NAME, which
# only our provisioning sets.
DEPLOYMENT_ID=
PROJECT_ID=
APP_KEY=

# -----------------------------------------------------------------------------
# Serverless Functions Configuration
# -----------------------------------------------------------------------------
# Deno Deploy - For hosting serverless edge functions
# Get your token from: https://dash.deno.com/account#access-tokens
# Get your org ID from: https://dash.deno.com/
DENO_DEPLOY_TOKEN=
DENO_DEPLOY_ORG_ID=

# Edge-function execution timeout in milliseconds.
# WORKER_TIMEOUT_MS=60000

# Edge-function host mode: development or production.
# DENO_ENV=production

# ─── Compute Services (Fly.io) ──────────────────────────────────────────────
# Deploy Docker containers with persistent URLs.
# Full setup + architecture: https://docs.insforge.dev/core-concepts/compute/architecture
#
# Setup — both FLY_API_TOKEN and FLY_ORG are required. FLY_ORG must be YOUR
# Fly org slug; leaving it blank or as "insforge" will fail with an opaque
# Fly auth error because that's our internal org. Compute auto-enables when
# both are set.
#   1. Create a Fly.io account: https://fly.io
#   2. Find your org slug: fly orgs list
#   3. Create an API token scoped to that org: fly tokens create org -o <your-org>
#   4. Set FLY_API_TOKEN=<token> and FLY_ORG=<your-org>
#
# Optional: Set COMPUTE_DOMAIN only if you own a wildcard domain pointed at
# Fly (e.g. *.compute.yourdomain.com CNAME'd to fly.dev). Leave unset to use
# Fly's own .fly.dev hostname, which works out of the box.

FLY_API_TOKEN=
FLY_ORG=
COMPUTE_DOMAIN=

# ─── Custom Compute: Docker provider (self-host, opt-in) ───────────────────────
# Runs your containers on THIS host instead of Fly.io. Enabled by mounting the
# Docker socket into the InsForge container — see the commented lines in
# docker-compose.prod.yml. The driver registers itself when the socket is
# reachable, so nothing runs containers on your host unless you grant that.
#
# Security: the provider builds every container spec itself and never forwards
# caller-supplied options, so a leaked InsForge API key cannot ask for a
# privileged container or a host bind mount. The socket itself is still
# root-equivalent on the host, so treat mounting it as the decision it is.
#
# Known limitation: there are no persistent volumes yet (the Fly path has none
# either). Container state survives restarts and host reboots, but changing the
# image, env vars, or port recreates the container and discards anything written
# inside it. Use the project's Postgres or Storage for data you need to keep.

# Which provider new services go to: fly | cloud | docker | off. Leave empty to
# auto-detect. Not a restriction — providers coexist, and each service is managed
# by the one that created it.
COMPUTE_PROVIDER=

# Ingress for services that do not choose their own:
#   none (default) — internal network only; no host port, no URL
#   port           — published on a host port
#   host           — reachable at a hostname via your own gateway
COMPUTE_DEFAULT_INGRESS=none

# Host address used to build URLs for `port` ingress. Left empty, no URL is
# advertised — better than handing out one that does not resolve.
COMPUTE_PUBLIC_HOST=

# Bind address for published ports. Loopback by default: Docker's own default
# publishes on 0.0.0.0 AND [::], i.e. the whole internet on a reachable host.
COMPUTE_BIND_ADDRESS=127.0.0.1

# Path to the Docker socket, if it is not in the usual place — rootless Docker
# ($XDG_RUNTIME_DIR/docker.sock) or Podman (/run/podman/podman.sock). The compose
# files mount it at this same path inside the container, so one value covers both
# sides. Only read when the socket mount is uncommented.
DOCKER_SOCKET_PATH=/var/run/docker.sock

# Ceiling on an uploaded source-build context (default 64mb). The whole tarball is
# buffered in memory before the build starts, and only one upload is accepted at a
# time, so this is the memory the build path can cost — lower it on a small host.
COMPUTE_BUILD_MAX_CONTEXT=

# Seconds a build upload may send nothing before it is treated as stalled and cut
# loose (default 30). Builds run one at a time, so a connection that stops making
# progress would otherwise block every other deploy. The timer resets on each chunk,
# so this bounds silence, not total upload time — a slow but active link is fine.
COMPUTE_BUILD_UPLOAD_IDLE_TIMEOUT=

# Set true to keep compute containers off this project's network. Off by default:
# sitting next to the database and storage is the point, and `postgres:5432`
# resolves from inside a compute container exactly as it does for edge functions.
COMPUTE_ISOLATE_NETWORK=

# Cloud-mode compute (provisioned automatically when PROJECT_ID + CLOUD_API_HOST
# are set by insforge-cloud). Self-host users with FLY_API_TOKEN always take
# priority and do not need to set PROJECT_ID.

# =============================================================================
# End of Configuration
# =============================================================================
# After configuring this file:
# 1. Run: docker compose up -d
# 2. Check logs: docker compose logs -f
# 3. Access the dashboard at: http://localhost:7131
#
# For more information, visit: https://insforge.dev/docs
# =============================================================================