> ## Documentation Index
> Fetch the complete documentation index at: https://docs.brainstormer.io/llms.txt
> Use this file to discover all available pages before exploring further.
# Environment & Configuration
> Environment variables, API keys, service-key mapping, agent configuration variables, and platform config reference.
## Required API Keys
These keys are required for core functionality. The platform will not work without them.
| Key | Purpose | Get It At |
| -------------------- | ------------------------------------------------------- | ---------------------------------------------------------------------------- |
| `OPENROUTER_API_KEY` | Access to 300+ AI models (GPT-4, Claude, Gemini, Llama) | [https://openrouter.ai/keys](https://openrouter.ai/keys) |
| `OPENAI_API_KEY` | Embeddings for Knowledge Base System | [https://platform.openai.com/api-keys](https://platform.openai.com/api-keys) |
## Optional API Keys
| Key | Purpose | Get It At |
| --------------------- | --------------------------------------------------- | ---------------------------------------------------------------------------------------------------------- |
| `ANTHROPIC_API_KEY` | Direct Claude access (alternative to OpenRouter) | [https://console.anthropic.com/](https://console.anthropic.com/) |
| `ELEVENLABS_API_KEY` | Voice cloning and text-to-speech | [https://elevenlabs.io/app/speech-synthesis/api-keys](https://elevenlabs.io/app/speech-synthesis/api-keys) |
| `LLAMA_CLOUD_API_KEY` | Advanced document parsing via LlamaParse | [https://cloud.llamaindex.ai/api-key](https://cloud.llamaindex.ai/api-key) |
| `PINECONE_API_KEY` | Production vector database (falls back to ChromaDB) | [https://www.pinecone.io/](https://www.pinecone.io/) |
| `GEMINI_API_KEY` | Gemini multimodal embeddings | [https://aistudio.google.com/apikey](https://aistudio.google.com/apikey) |
## Service-Key Mapping
| Service | Port | Required Keys | Optional Keys |
| --------------------- | ---- | -------------------- | ----------------------------------------------------------- |
| **Bot Service** | 4002 | `OPENROUTER_API_KEY` | `ANTHROPIC_API_KEY` |
| **Knowledge Service** | 4005 | `OPENAI_API_KEY` | `LLAMA_CLOUD_API_KEY`, `PINECONE_API_KEY`, `GEMINI_API_KEY` |
| **Realtime Audio** | 4003 | None | `ELEVENLABS_API_KEY` |
| **Auth Service** | 4001 | None | `RESEND_API_KEY` |
| **Gateway** | 4000 | None | None |
| **Web App** | 3000 | None | None |
## Full Environment Reference
```bash theme={null}
# PostgreSQL (Docker container)
DATABASE_URL=postgresql://brainstormer:password@localhost:5432/brainstormer
# Redis (Docker container)
REDIS_HOST=localhost
REDIS_PORT=6379
REDIS_PASSWORD=
# ChromaDB
CHROMA_HOST=localhost
CHROMA_PORT=8000
```
```bash theme={null}
# OpenRouter (REQUIRED)
OPENROUTER_API_KEY=sk-or-v1-...
OPENROUTER_BASE_URL=https://openrouter.ai/api/v1
# OpenAI (REQUIRED for embeddings)
OPENAI_API_KEY=sk-...
# Gemini Embeddings
GEMINI_API_KEY=...
EMBEDDING_MODEL=gemini-embedding-exp-03-07
EMBEDDING_DIMENSIONS=3072
# Default AI model
DEFAULT_MODEL=google/gemini-2.0-flash-001
PLATFORM_INTELLIGENCE_MODEL=anthropic/claude-3-haiku
```
```bash theme={null}
GATEWAY_PORT=4000
AUTH_SERVICE_PORT=4001
BOT_SERVICE_PORT=4002
REALTIME_AUDIO_PORT=4003
KNOWLEDGE_SERVICE_PORT=4005
```
```bash theme={null}
JWT_SECRET=your-super-secret-jwt-key
JWT_EXPIRES_IN=24h
JWT_REFRESH_EXPIRES_IN=7d
BCRYPT_ROUNDS=12
# Superadmin access
PLATFORM_SUPERADMIN_EMAILS=admin@brainstormer.ai
NEXT_PUBLIC_PLATFORM_SUPERADMIN_EMAILS=admin@brainstormer.ai
```
```bash theme={null}
RESEND_API_KEY=re_...
EMAIL_FROM=noreply@brainstormer.ai
WEB_APP_URL=https://bsio2.brainstormer.io
```
```bash theme={null}
CLOUD_STORAGE_ENABLED=false
CLOUD_STORAGE_PROVIDER=aws-s3
CLOUD_STORAGE_BUCKET=your-bucket-name
CLOUD_STORAGE_REGION=us-east-1
CLOUD_STORAGE_ACCESS_KEY_ID=...
CLOUD_STORAGE_SECRET_ACCESS_KEY=...
```
```bash theme={null}
# Pinecone (production)
PINECONE_API_KEY=...
PINECONE_ENVIRONMENT=us-east-1
PINECONE_INDEX_NAME=brainstormer-kb
# Processing
CHUNK_SIZE=1000
CHUNK_OVERLAP=200
MAX_CONCURRENT_JOBS=5
```
```bash theme={null}
ELEVENLABS_API_KEY=...
LIVEKIT_API_KEY=...
LIVEKIT_API_SECRET=...
LIVEKIT_URL=wss://...
```
```bash theme={null}
LANGSMITH_TRACING=true
LANGSMITH_API_KEY=...
LANGSMITH_PROJECT=brainstormer
MODEL_SYNC_ENABLED=true
MODEL_SYNC_INTERVAL_MINUTES=60
```
## Agent Configuration Variables
These frontend-facing configuration variables control agent creation and validation:
### Agent Name Validation
| Variable | Default | Description |
| --------------------------------- | ------- | ------------------------- |
| `NEXT_PUBLIC_BOT_NAME_MIN_LENGTH` | 3 | Minimum agent name length |
| `NEXT_PUBLIC_BOT_NAME_MAX_LENGTH` | 50 | Maximum agent name length |
### Agent Prompt Validation
| Variable | Default | Description |
| ------------------------------------------ | ------- | ------------------------ |
| `NEXT_PUBLIC_BOT_SYSTEM_PROMPT_MAX_LENGTH` | 1000 | Max system prompt length |
| `NEXT_PUBLIC_BOT_DESCRIPTION_MAX_LENGTH` | 200 | Max description length |
### Agent UI
| Variable | Default | Description |
| ------------------------------------ | ------- | ---------------------------- |
| `NEXT_PUBLIC_BOT_MAX_MODELS_DISPLAY` | 8 | Models shown in selection UI |
### Template Default Models
| Variable | Default |
| ------------------------------------------------- | --------------------------- |
| `NEXT_PUBLIC_BOT_TEMPLATE_CUSTOMER_SUPPORT_MODEL` | `anthropic/claude-3-haiku` |
| `NEXT_PUBLIC_BOT_TEMPLATE_KNOWLEDGE_MODEL` | `openai/gpt-4o-mini` |
| `NEXT_PUBLIC_BOT_TEMPLATE_CREATIVE_MODEL` | `anthropic/claude-3-sonnet` |
| `NEXT_PUBLIC_BOT_TEMPLATE_CODING_MODEL` | `openai/gpt-4o` |
Configuration keys maintain `BOT_` prefix for backward compatibility during transition from "bots" to "agents" terminology.
## What Won't Work Without Keys
| Missing Key | Impact |
| --------------------- | -------------------------------------------------------------------------- |
| `OPENROUTER_API_KEY` | Cannot chat with AI agents, model sync disabled, agent testing unavailable |
| `OPENAI_API_KEY` | Knowledge Base embeddings won't work, document upload/search disabled |
| `ELEVENLABS_API_KEY` | Voice cloning features disabled |
| `LLAMA_CLOUD_API_KEY` | Falls back to open-source parsers for document processing |
| `PINECONE_API_KEY` | Falls back to ChromaDB (local) for vector storage |
## Cost Estimates (Monthly)
### Required (Minimal Usage)
| Service | Cost |
| ----------------- | ----------------------------------------------- |
| OpenRouter | \$0-5/month (pay per use, includes free models) |
| OpenAI Embeddings | \~\$0.10 per 1M tokens |
### Optional
| Service | Cost |
| ---------- | --------------------------------- |
| Anthropic | \$0-20/month |
| ElevenLabs | Free tier (10k chars/month) |
| LlamaParse | Free tier (1k pages/day) |
| Pinecone | Free tier (1 index, 100k vectors) |
## Platform Config (DB-Stored)
In addition to environment variables, API keys and settings can be stored in the `platform_config` database table, managed via the admin UI at `/admin`. DB values take precedence over env vars.
**Categories:** `ai`, `voice`, `knowledge`, `feature_flags`
**Managed keys include:** OPENROUTER, OPENAI, ELEVENLABS, LIVEKIT, PINECONE, GEMINI, and more.
## Security Notes
* Never commit `.env` to git (already in `.gitignore`)
* Use different keys for development and production
* Rotate keys regularly if exposed
* Monitor usage on provider dashboards
* Sensitive platform config values are encrypted with AES-256-GCM