otx implement
Generate working code from a compose or transmute output spec file.
Synopsis
otx implement <spec> [options]
Arguments
| Argument | Description |
spec | Path to the spec file (e.g. Output/Composition/my-design/compose-output.md) (required) |
Options
| Option | Default | Description |
--output, -o | Output/Implementation/<project> | Output directory for generated code |
--project, -p | inferred from spec path | Project name passed to the model for correct naming |
--orchestrator | from settings | ClaudeCode | OpenAI | Ollama |
--api-key | OPENAI_API_KEY env | OpenAI API key (never saved to disk) |
--endpoint | from settings | Custom OpenAI-compatible base URL |
--model | thick model from settings | Model override (defaults to the most capable/thick tier) |
--max-turns | 200 | Max agent turns |
--timeout | 60 | HTTP timeout in minutes |
Examples
# Basic usage — implement a compose output
otx implement Output/Composition/my-design/compose-output.md
# Explicit project name and output directory
otx implement Output/Composition/my-design/compose-output.md \
--project my-new-system \
--output ./src/my-new-system
# With a specific backend and model
otx implement path/to/spec.md \
--orchestrator OpenAI \
--api-key sk-... \
--model gpt-4o \
--max-turns 300 \
--timeout 120
Notes
- The output directory is not cleared before a run. The agent will see any files already present and can build on or correct them.
- ClaudeCode is recommended — it uses its own file tools and handles large multi-file implementations naturally.
- For complex specs, increase
--max-turns (try 300–500) and --timeout (try 90–120 minutes).