Commands Reference¶
Complete reference for all clinvk CLI commands.
Synopsis¶
Command Overview¶
| Command | Description | Common Use |
|---|---|---|
[prompt] |
Execute a prompt (default command) | Daily AI tasks |
resume |
Resume a previous session | Continue conversations |
sessions |
Manage sessions | List, show, delete sessions |
config |
Manage configuration | View or change settings |
parallel |
Execute tasks in parallel | Run multiple tasks |
compare |
Compare backend responses | Evaluate different AIs |
chain |
Execute prompt chain | Multi-step workflows |
mcp |
Start MCP server | MCP integrations |
serve |
Start HTTP API server | Application integration |
version |
Show version information | Check installed version |
help |
Show help | Get command help |
Global Flags¶
These flags work with all commands:
| Flag | Short | Type | Default | Description |
|---|---|---|---|---|
--backend |
-b |
string | claude |
AI backend to use |
--model |
-m |
string | Model to use | |
--workdir |
-w |
string | Working directory | |
--output-format |
-o |
string | json |
Output format |
--config |
string | Config file path | ||
--dry-run |
bool | false |
Show command only | |
--ephemeral |
bool | false |
Stateless mode | |
--help |
-h |
Show help |
Prompt-Specific Flags¶
These flags work only with the default prompt command:
| Flag | Short | Type | Default | Description |
|---|---|---|---|---|
--continue |
-c |
bool | false |
Continue the most recent session |
Flag Details¶
--backend, -b¶
Select the AI backend to use:
Available backends: claude, codex, gemini
--model, -m¶
Override the default model for the selected backend:
--workdir, -w¶
Set the working directory for the AI backend:
--output-format, -o¶
Control output format:
| Value | Description |
|---|---|
text |
Plain text |
json |
Structured JSON (default) |
stream-json |
Streaming JSON events |
--config¶
Use a custom configuration file:
--dry-run¶
Show the command that would be executed without running it:
clinvk --dry-run "implement feature X"
# Output: Would execute: claude --model claude-opus-4-5-20251101 "implement feature X"
--ephemeral¶
Run in stateless mode without creating a session:
Command Categories¶
Core Commands¶
Commands for everyday use:
[prompt]- Execute promptsresume- Continue sessionssessions- Manage sessions
Configuration Commands¶
Commands for managing settings:
config- View and modify configuration
Execution Commands¶
Commands for advanced execution patterns:
parallel- Run multiple tasks concurrentlychain- Execute sequential pipelinescompare- Compare multiple backends
Server Commands¶
Commands for running the HTTP API:
mcp- Start the MCP serverserve- Start the API server
Usage Examples¶
Basic Usage¶
# Execute a prompt
clinvk "fix the bug in auth.go"
# Specify backend
clinvk -b codex "implement feature"
# Specify model
clinvk -b claude -m claude-sonnet-4-20250514 "quick review"
Session Management¶
# List sessions
clinvk sessions list
# Show session details
clinvk sessions show abc123
# Resume last session
clinvk resume --last
# Delete old sessions
clinvk sessions clean --older-than 30d
Configuration¶
Advanced Execution¶
# Run tasks in parallel
clinvk parallel --file tasks.json
# Compare backends
clinvk compare --all-backends "explain this code"
# Execute a chain
clinvk chain --file pipeline.json
Server¶
Exit Codes¶
All commands return exit codes:
| Code | Description |
|---|---|
| 0 | Success |
| 1 | General error |
| 2 | Backend not available |
| 3 | Invalid configuration |
| 4 | Session error |
See Exit Codes for complete reference.
Getting Help¶
Get help for any command:
# General help
clinvk --help
# Command help
clinvk [command] --help
# Example
clinvk parallel --help
See Also¶
- Configuration Reference - Configuration options
- Environment Variables - Environment-based settings
- Exit Codes - Exit code reference