Quickstart¶
Get productive with DevDuck in 2 minutes.
Interactive REPL¶
🦆 DevDuck
📝 Logs: /tmp/devduck/logs
🦆 ✓ Zenoh peer: hostname-abc123
🦆 ✓ WebSocket server: localhost:10001
🦆 create a python web server with FastAPI
Type your query and DevDuck handles the rest — writing files, running commands, installing packages.
One-Shot Mode¶
devduck "explain this error" < error.log
devduck "create a REST API for user management"
devduck "review the last 5 commits on this repo"
Multi-Conversation TUI¶
Run multiple conversations concurrently in a Textual-based terminal UI. Each conversation has its own panel with streaming markdown output.
Python API¶
import devduck
# Direct call
devduck("analyze this code and suggest improvements")
# Or use the ask helper
from devduck import ask
result = ask("what files are in this directory?")
print(result)
Shell Commands¶
Prefix with ! to run shell commands directly:
Key Workflows¶
Load a Tool at Runtime¶
DevDuck will install the package and load the tool — no restart needed.
Fetch a Tool from GitHub¶
🦆 fetch the scraper tool from https://github.com/cagataycali/devduck/blob/main/devduck/tools/scraper.py
Record a Session¶
🦆 record
... do your work ...
🦆 record
🎬 Session exported: /tmp/devduck/recordings/session-20260326.zip
Toggle Ambient Mode¶
Configuration¶
Key environment variables:
| Variable | Default | Description |
|---|---|---|
MODEL_PROVIDER |
auto-detect | Force model provider |
STRANDS_MODEL_ID |
provider default | Specific model ID |
DEVDUCK_TOOLS |
core set | Tool configuration |
DEVDUCK_AMBIENT_MODE |
false |
Enable ambient mode |
DEVDUCK_ENABLE_WS |
true |
WebSocket server |
DEVDUCK_ENABLE_ZENOH |
true |
Zenoh P2P |
MCP_SERVERS |
— | MCP server config (JSON) |
Next Steps¶
- Hot Reload — Edit code while the agent runs
- Tools — Load, create, and manage tools
- Models — All 14 supported providers
- Zenoh P2P — Multi-instance networking