NanoClaw uses a zero-config-file approach. Configuration is handled through Claude Code skills during setup, with runtime behavior controlled by the application’s built-in defaults and per-group settings.
NanoClaw intentionally has 0 configuration files. Instead:
/setup skill/add-telegram) rather than config changesDuring setup, you’ll choose your container runtime:
| Platform | Default Runtime | Alternative |
|---|---|---|
| macOS | Apple Container | Docker |
| Linux | Docker | - |
| Windows | WSL2 + Docker | - |
Each agent/chat group gets:
Concurrency defaults:
By default, all shell commands run inside isolated containers:
Configured during /setup:
Add channels by running these commands in Claude Code:
/add-telegram # Add Telegram support
/add-slack # Add Slack support
/add-discord # Add Discord support
/add-signal # Add Signal support
Each skill configures its own authentication and connection settings.
NanoClaw supports Agent Swarms via the Claude Agent SDK. Configure swarm behavior:
Example swarm roles:
- Planner: Breaks down complex tasks
- Executor: Performs actions and runs commands
- Reviewer: Validates results before responding
- Researcher: Fetches web information
Configure recurring tasks via Claude Code:
# Morning briefing (daily at 8 AM)
# Weekly review (every Monday)
# Custom cron expressions
Tasks are stored in SQLite and executed by the task scheduler.
Default data locations:
| Data Type | Default Location |
|---|---|
| SQLite Database | ./data/nanoclaw.db (relative to install) |
| Group Memory (CLAUDE.md) | ./groups/<group-id>/CLAUDE.md |
| Container Mounts | Per-group isolated directories |
| Session Data | SQLite database |
Back up the following:
nanoclaw.db)./groups/) with CLAUDE.md filesRecovery process:
/setup⚠️ Note: NanoClaw does not use traditional environment variables for configuration. The following are speculative examples for reference only. Actual configuration is via Claude Code skills:
# These are EXAMPLE/PLACEHOLDER values only
# NanoClaw uses Claude Code /setup for actual configuration
NANOCLAW_BIND_ADDR=127.0.0.1:3000 # Speculative
NANOCLAW_DATA_DIR=/var/lib/nanoclaw # Speculative
NANOCLAW_LOG_LEVEL=info # Speculative
NANOCLAW_SAFE_BASH=true # Enabled by default via container isolation
NANOCLAW_CONTAINER_RUNTIME=docker # Set during /setup
NANOCLAW_SWARM_DEFAULT=planner,executor,reviewer # Configured per-group
Any questions?
Feel free to contact us. Find all contact information on our contact page.