Configuration
Configure the Hybrix gateway, agents, skills, and channels through a single JSON configuration file.
Configuration file
Hybrix reads its configuration from hybrix.config.json in your project root. The onboarding wizard creates this file for you automatically.
json
{
"gateway": {
"port": 3000,
"host": "127.0.0.1",
"routing": "role-based",
"cors": ["http://localhost:3000"]
},
"agents": [],
"skills": [],
"channels": []
}Gateway settings
| Option | Default | Description |
|---|---|---|
| port | 3000 | HTTP port for the gateway and dashboard |
| host | 127.0.0.1 | Bind address. Use 0.0.0.0 for remote access |
| routing | role-based | Routing strategy: role-based, round-robin, or intent |
| cors | [] | Allowed CORS origins |
| maxConcurrent | 10 | Max concurrent agent conversations |
Agent configuration
Each agent in the agents array defines an AI persona with its own LLM provider, role, and skill set.
json
{
"agents": [
{
"name": "Atlas",
"role": "general-assistant",
"provider": "anthropic",
"model": "claude-sonnet-4-20250514",
"systemPrompt": "You are Atlas, a helpful general assistant.",
"skills": ["shell-exec", "web-browse"],
"channels": ["discord", "web"],
"temperature": 0.7,
"maxTokens": 4096
}
]
}Each agent can use a different LLM provider. This lets you route cost-sensitive queries to cheaper models and complex tasks to more capable ones.
Environment variables
API keys and secrets should be stored in a .env file, not in the config:
env
ANTHROPIC_API_KEY=sk-ant-...
OPENAI_API_KEY=sk-...
GOOGLE_AI_KEY=...
DISCORD_BOT_TOKEN=...
SLACK_BOT_TOKEN=xoxb-...Never commit your
.env file to version control. Add it to your .gitignore.