Config commands

Reference for every indexed config * subcommand (see indexed config --help), configuration keys, and environment overrides. For global indexed options, see Other commands — global options (top-level). For narrative setup and tuning, see the Configuration guide.

`indexed config init`

Create a workspace .indexed/ directory in the current working directory with config.toml and .env.example.

indexed config init [flags]

Flag	Type	Required	Default	Description
`--force` / `-f`	flag	No	off	Overwrite existing workspace config files

Logging options (--verbose, --log-level, --json-logs) apply to this command.

Example:

cd ~
indexed config init

This creates ~/.indexed/ when your shell is in your home directory (because .indexed is resolved relative to the cwd). To use a project-local workspace instead, run indexed config init from your repository root — then config lives at <repo>/.indexed/.

`indexed config inspect`

Display merged configuration (global + workspace + environment).

indexed config inspect [section] [flags]

Argument / flag	Type	Required	Default	Description
`section`	string	No	—	Optional: `sources`, `core`, `logging`, `mcp`, `performance`
`--defaults` / `-d` / `--show-defaults`	flag	No	off	Show all values including defaults

Examples:

indexed config inspect
indexed config inspect sources
indexed config inspect --defaults

JSON: use the global --simple-output flag:

indexed --simple-output config inspect

`indexed config set`

Set a value in the workspace config.toml (dot path under .indexed/ for the workspace you initialized).

indexed config set <key> <value> [flags]

Flag	Type	Required	Default	Description
`key`	string	Yes	—	Config key path (e.g., `core.v1.indexing.chunk_size`)
`value`	string	Yes	—	New value (auto-coerced)
`--dry-run`	flag	No	off	Preview the change without saving

Example:

indexed config set core.v1.indexing.chunk_size 1024
indexed config set core.v1.indexing.chunk_size 1024 --dry-run

`indexed config update`

Update global configuration only — interactively, or replace it entirely from a TOML file. The --file path replaces the global config, not the workspace .indexed/config.toml from config init.

indexed config update [flags]

Flag	Type	Required	Default	Description
`--file` / `-f`	path	No	—	Path to a TOML file that replaces the global configuration

Examples:

indexed config update
indexed config update --file ./my-global-config.toml

`indexed config delete`

Remove a key from the workspace configuration (prompts for confirmation unless --force).

indexed config delete <key> [flags]

Flag	Type	Required	Default	Description
`--force` / `-f`	flag	No	off	Skip confirmation

Example:

indexed config delete core.v1.indexing.chunk_overlap
indexed config delete core.v1.indexing.chunk_overlap --force

`indexed config validate`

Validate the active configuration against registered validation rules. Prints errors by section and exits with status 1 if validation fails.

indexed config validate

Run this after manually editing workspace or global config.toml (see the Configuration guide) to catch typos before they cause indexing errors.

Configuration Keys

Indexing

Key	Type	Default	Description
`core.v1.indexing.chunk_size`	int	`512`	Maximum number of tokens per chunk
`core.v1.indexing.chunk_overlap`	int	`50`	Number of overlapping tokens between consecutive chunks

Embedding

Key	Type	Default	Description
`core.v1.embedding.model_name`	string	`"all-MiniLM-L6-v2"`	Sentence Transformers model to use for embeddings

Files connector

Keys below belong in a [sources.files] table in config.toml. With indexed config set, use the full dot path (for example, sources.files.ocr_enabled true).

Key	Type	Default	Description
`sources.files.path`	string	`"."`	Root directory to index (`-p` / `--path` overrides for a single `index create` run)
`sources.files.include_patterns`	list	`["*"]`	Glob-style patterns for paths to include (for example `.md`, `/.py`)
`sources.files.exclude_patterns`	list	(built-in noise paths)	Regex patterns matched against full paths; defaults skip common dirs such as `.git`, `node_modules`, and `.venv` (see `indexed config inspect --defaults`)
`sources.files.fail_fast`	bool	`false`	Stop indexing on the first file read error
`sources.files.respect_gitignore`	bool	`true`	Honor `.gitignore` and skip common noise directories
`sources.files.change_tracking`	string	`"auto"`	Change detection: `auto`, `git`, `content_hash`, `mtime`, or `none`
`sources.files.ocr_enabled`	bool	`false`	Enable OCR for scanned PDFs and images
`sources.files.table_structure`	bool	`true`	Extract table structure from documents
`sources.files.code_chunking`	bool	`true`	Use AST-aware chunking for supported code languages
`sources.files.max_chunk_tokens`	int	`512`	Maximum tokens per chunk in the parsing module
`sources.files.excluded_extensions`	list	`[]`	Extensions to skip (e.g. `[".lock", ".min.js"]`)

Precedence

Merged configuration follows global + workspace + environment (indexed config inspect --help). From lowest to highest priority:

Built-in defaults
~/.indexed/config.toml (global), when present
.indexed/config.toml for your workspace (overrides global where set)
Environment variables (INDEXED__… and connector secrets)
CLI flags on the current command

Environment Variable Overrides

Any configuration key can be overridden with an environment variable using the pattern:

INDEXED__<section>__<subsection>__<key>=<value>

Double underscores (__) separate each level of the config hierarchy. Environment variables override values from config files.

Examples:

export INDEXED__core__v1__indexing__chunk_size=1024
export INDEXED__core__v1__indexing__chunk_overlap=100
export INDEXED__core__v1__embedding__model_name="paraphrase-multilingual-MiniLM-L12-v2"

Connector Credentials

Credentials are read from environment variables only — never from config.toml. CLI flags for indexed index create are in Index commands; this section lists the env vars.

Variable	Used by	Description
`ATLASSIAN_EMAIL`	Jira Cloud, Confluence Cloud	Atlassian account email
`ATLASSIAN_TOKEN`	Jira Cloud, Confluence Cloud	Atlassian API token
`JIRA_TOKEN`	Jira Server/DC	Bearer or personal access token
`JIRA_LOGIN`	Jira Server/DC	Username (with `JIRA_PASSWORD`)
`JIRA_PASSWORD`	Jira Server/DC	Password (with `JIRA_LOGIN`)
`CONF_TOKEN`	Confluence Server/DC	Bearer or personal access token
`CONF_LOGIN`	Confluence Server/DC	Username (with `CONF_PASSWORD`)
`CONF_PASSWORD`	Confluence Server/DC	Password (with `CONF_LOGIN`)

See Jira and Confluence for setup, Cloud vs Server/DC, and troubleshooting.

Keep secrets out of version control

Use .indexed/.env (or your shell) for secrets — not config.toml. Add .env to .gitignore if you keep one in a repo directory.

Example Config File

~/.indexed/config.toml

[core.v1.indexing]
chunk_size = 512
chunk_overlap = 50

[core.v1.embedding]
model_name = "all-MiniLM-L6-v2"

[sources.files]
path = "./documents"
include_patterns = ["*.md", "*.pdf"]
exclude_patterns = ["\\.draft\\.md$", "/build/"]
respect_gitignore = true
change_tracking = "auto"
ocr_enabled = false

Data directory

Indexed stores data under ~/.indexed/:

~/.indexed/
├── config.toml
└── data/
    └── collections/<name>/
        ├── manifest.json
        ├── documents.json
        ├── chunks.json
        └── index.faiss

See What is Indexed for how local data is used.

Config commands

On this page