API Documentation
gitaiflow exposes its API through the gitaiflow executable — a compiled native binary. There is no HTTP API and no subcommand structure; every feature is a flag on the single root invocation.
On this page ▾
Entry point
gitaiflow --helpCurrent command surface
gitaiflow [--path PATH] [--remote REMOTE] [--base-branch BASE_BRANCH]
[--skip [PATH_OR_PATTERN ...]] [-o OUTPUT_DIR] [--markdown]
[--no-chunk] [--change-summary] [--show-analysis]
[--last-summary] [--regenerate-summary] [--allow-billing]
[--changelog] [--since DATE] [--until DATE] [--release-notes]
[--list-models] [--free-only] [--json]
[--version] [--usage [current|VERSION]]
[--telemetry {enable,disable,status,history}] [--uninstall]Execution modes
gitaiflow has four independent execution modes, selected by which flags are present:
| Mode | Trigger | Calls an AI provider? | Touches Git? |
|---|---|---|---|
| Normal summary | default (no mode flag) | Yes | Yes |
| Changelog | --changelog |
No | Yes (or reads existing JSON artifacts) |
| Model catalog | --list-models |
No (queries OpenRouter's public catalog) | No |
| Usage / telemetry | --usage, --telemetry |
No | No |
--list-models, --usage, --telemetry, and --version exit immediately and never run the summary pipeline.
Global options
gitaiflow --version # print version and exit
gitaiflow --help # print usage and exitFeature flags
Target selection
gitaiflow --path mailer/Provider-required summary generation
gitaiflow --path . --change-summaryRequires a configured AI provider — see configuration.md. Fails fast with configuration instructions rather than partway through a run if nothing is configured.
Changelog generation
gitaiflow --changelog --since 2026/08/01 --path .Does not call an AI provider.
Model discovery
gitaiflow --list-models --free-onlyIndependent of the configured AI_PROVIDER — always targets OpenRouter's public catalog endpoint, and doesn't require an API key.
Telemetry consent
gitaiflow --telemetry statusExit behavior
| Condition | Behavior |
|---|---|
| Successful summary / changelog / model-list / usage / telemetry command | Exit 0 |
| No AI provider configured (normal summary mode) | Fails fast with configuration instructions before any AI request is attempted |
--changelog range resolves entirely before repository history, or into the future |
Rejected upfront with an error, no changelog written |
| AI provider call fails | Non-fatal to the changelog/model-catalog/usage/telemetry paths; a normal summary run reports the failure |
Full command reference
See ../user-guide/commands.md for every flag with usage examples and output schema, and configuration.md for the environment-variable contract that backs provider selection.