Configuration

Erode is configured through environment variables or a .eroderc.json configuration file. For GitHub Actions-specific inputs (model-repo, fail-on-violations, etc.), see GitHub Actions.

Precedence

Erode resolves configuration in the following order (highest priority wins):

Environment variables (ERODE_*), whether set directly, through CI secrets, or via a .env file
.eroderc.json, project-level settings from the config file
Defaults, built-in defaults for all settings

This means you can commit .eroderc.json with project settings (model format, constraints, adapter config) and provide secrets (API keys, tokens) through environment variables in CI or a local .env file.

Configuration file

Create a .eroderc.json file in your project root or home directory:

{
  "$schema": "https://erode.dev/schemas/v0/eroderc.schema.json",
  "ai": { "provider": "gemini" }
}

The $schema field enables autocomplete and validation in editors that support JSON Schema. All sections are optional; Erode fills in defaults for anything you omit.

Erode looks for .eroderc.json in the current working directory first, then the home directory. Any ERODE_* environment variables override matching values from the config file.

Environment variable names map to nested JSON keys. For example:

Environment variable	JSON path
`ERODE_AI_PROVIDER`	`ai.provider`
`ERODE_GEMINI_API_KEY`	`gemini.apiKey`
`ERODE_MAX_FILES_PER_DIFF`	`constraints.maxFilesPerDiff`
`ERODE_MODEL_FORMAT`	`adapter.format`
`ERODE_MODEL_PATH`	`adapter.modelPath`
`ERODE_MODEL_REPO`	`adapter.modelRepo`
`ERODE_MODEL_REF`	`adapter.modelRef`

AI provider

Variable	Description	Default
`ERODE_AI_PROVIDER`	AI provider to use (`gemini`, `openai`, or `anthropic`)	`gemini`
`ERODE_GEMINI_API_KEY`	Google Gemini API key	—
`ERODE_OPENAI_API_KEY`	OpenAI API key	—
`ERODE_ANTHROPIC_API_KEY`	Anthropic API key (experimental)	—

Architecture model

Variable	Description	Default
`ERODE_MODEL_FORMAT`	Architecture model format (`likec4` or `structurizr`)	`likec4`
`ERODE_MODEL_PATH`	Path to the architecture model directory	—
`ERODE_MODEL_REPO`	Repository URL or owner/repo containing the model	—
`ERODE_MODEL_REF`	Branch or tag to clone from the model repository	`main`
`ERODE_STRUCTURIZR_CLI_PATH`	Path to the Structurizr CLI WAR file (for `.dsl` files)	—
`ERODE_LIKEC4_EXCLUDE_PATHS`	Comma-separated paths to exclude from model loading	—
`ERODE_LIKEC4_EXCLUDE_TAGS`	Comma-separated tags to exclude from model loading	—

Diff limits

Variable	Description	Default
`ERODE_MAX_FILES_PER_DIFF`	Maximum number of files to include in the diff	`50`
`ERODE_MAX_LINES_PER_DIFF`	Maximum number of lines to include in the diff	`5000`
`ERODE_MAX_CONTEXT_CHARS`	Maximum characters of architectural context	`10000`

Erode truncates large diffs to stay within these limits. If a PR exceeds them, Erode processes the most relevant files first based on the architecture model context.

Provider model overrides

Each AI provider uses two model tiers: a fast model for extraction stages and an advanced model for analysis. Override the defaults with these variables:

Variable	Description
`ERODE_GEMINI_FAST_MODEL`	Gemini model for Stages 1–2 and model updates
`ERODE_GEMINI_ADVANCED_MODEL`	Gemini model for Stage 3 (analysis)
`ERODE_OPENAI_FAST_MODEL`	OpenAI model for Stages 1–2 and model updates
`ERODE_OPENAI_ADVANCED_MODEL`	OpenAI model for Stage 3 (analysis)
`ERODE_ANTHROPIC_FAST_MODEL`	Anthropic model for Stages 1–2 and model updates
`ERODE_ANTHROPIC_ADVANCED_MODEL`	Anthropic model for Stage 3 (analysis)

See AI Providers for default model names and guidance on choosing a provider.

Platform tokens

Variable	Description	Default
`ERODE_GITHUB_TOKEN`	GitHub token for API access	—
`ERODE_GITHUB_BASE_URL`	GitHub API base URL	`https://api.github.com`
`ERODE_MODEL_REPO_PR_TOKEN`	Separate token for the model repository	Uses `ERODE_GITHUB_TOKEN`
`ERODE_GITLAB_TOKEN`	GitLab token with `api` scope	—
`ERODE_GITLAB_BASE_URL`	GitLab instance URL	`https://gitlab.com`
`ERODE_BITBUCKET_TOKEN`	Bitbucket app password or token	—
`ERODE_BITBUCKET_BASE_URL`	Bitbucket API base URL	`https://api.bitbucket.org/2.0`

See Authentication for required permissions, token types, and platform-specific setup.

Timeouts

Variable	Description	Default
`ERODE_GEMINI_TIMEOUT`	Request timeout for Gemini API calls (ms)	`60000`
`ERODE_OPENAI_TIMEOUT`	Request timeout for OpenAI API calls (ms)	`60000`
`ERODE_ANTHROPIC_TIMEOUT`	Request timeout for Anthropic API calls (ms)	`60000`
`ERODE_GITHUB_TIMEOUT`	Request timeout for GitHub API calls (ms)	`30000`

Debug

Variable	Description	Default
`ERODE_DEBUG_MODE`	Enable debug output	`false`
`ERODE_VERBOSE`	Enable verbose logging	`false`

What’s next

CLI Commands: run Erode locally against any pull request
Authentication: token permissions for GitHub, GitLab, and Bitbucket
AI Providers: supported providers, default models, and overrides
Analysis pipeline: stage-by-stage reference for the analysis engine