Configuration

Quaid reads mutable runtime defaults from three places, in order of precedence: command-line flags, environment variables, and the config table inside memory.db. Anything that fixes the database identity (most notably the embedding model and schema generation) is also recorded in the immutable quaid_config table at init time.

Environment variables

User-facing variables. Internal/test-only flags are intentionally omitted.

Variable	Used by	Default	Notes
QUAID_DB	every command	~/.quaid/memory.db	Path to the memory database file.
QUAID_MODEL	every command	small	Embedding model alias (small/base/large/m3) or any Hugging Face model ID. On the airgapped build, non-default values produce a warning and are ignored.
QUAID_CHANNEL	install script	airgapped	Selects which release channel the shell installer downloads. Set to online for the smaller binary that pulls model assets at first use.
QUAID_WATCH_DEBOUNCE_MS	watcher / vault sync	1500	Debounce window before the file watcher flushes a batch of changes. Higher values reduce churn during noisy edits.
QUAID_QUARANTINE_TTL_DAYS	quarantine GC	30	Auto-discard threshold for clean quarantined pages with no remaining DB-only state.
QUAID_RAW_IMPORTS_KEEP	raw-import GC	10	Per-page cap on retained raw_imports history rows.
QUAID_RAW_IMPORTS_TTL_DAYS	raw-import GC	90	Time-to-live for inactive raw_imports rows.
QUAID_RAW_IMPORTS_KEEP_ALL	raw-import GC	unset	Set to 1 to disable raw-import GC entirely.
QUAID_FULL_HASH_AUDIT_DAYS	collection audit	7	Interval after which the full-hash audit pass re-hashes a file.
QUAID_HF_BASE_URL	online-model fetch	https://huggingface.co	Override the Hugging Face origin (e.g. for an internal mirror).
QUAID_MODEL_CACHE_DIR	online-model fetch	~/.quaid/models/	Cache directory for downloaded model weights.

The --db and --model CLI flags are global and override the corresponding env vars on a per-invocation basis.

Filesystem layout

Quaid owns one directory under your home folder:

~/.quaid/
├── skills/                 # Skill overrides; run `quaid skills extract` to populate
│   ├── ingest/SKILL.md
│   ├── query/SKILL.md
│   └── …
└── models/                 # Cached model weights (online build only)
    └── BAAI__bge-base-en-v1.5/

Custom skills resolve in this order: working-directory SKILL.md → ~/.quaid/skills/<name>/SKILL.md → embedded default. Run quaid skills doctor to see what is shadowing what.

The runtime default memory path is ~/.quaid/memory.db. If Quaid cannot resolve a home directory, it falls back to memory.db in the current working directory. SQLite WAL sidecars (memory.db-wal, memory.db-shm) are created at runtime and folded back in by quaid compact.

Mutable runtime config

The config table inside memory.db holds runtime defaults editable via quaid config:

quaid config get default_token_budget
quaid config set default_token_budget 6000
quaid config reset default_token_budget

Recognized keys (with seeded defaults):

Key	Default	Effect
version	10	Schema version mirror, kept in sync with quaid_config.schema_version. Don’t edit by hand.
embedding_model	BAAI/bge-small-en-v1.5	Mirror of the active model from quaid_config. Read-only; managed by init.
embedding_dimensions	384	Mirror of the active model’s vector width.
chunk_strategy	section	How pages are split before embedding.
search_merge_strategy	set-union	How FTS5 and vector results are fused.
default_token_budget	4000	Used by query --depth auto.
graph_depth	0	Hops walked during hybrid_search graph expansion (0 disables it). The --hops N CLI flag overrides this per-invocation.
graph_distance_decay	0.5	Multiplied per hop into expanded-candidate scores.
graph_expansion_max	50	Per-query cap on newly added graph candidates.
edge_weight_frontmatter	1.0	Default weight on derived frontmatter link rows.
edge_weight_entity_pattern	0.7	Reserved for the entity-pattern edge follow-on; not yet used as a links source.
edge_weight_wikilink	0.5	Default weight on derived wiki_link link rows.

Immutable model config

quaid_config holds metadata that cannot change after init without rebuilding the database:

Key	Example	Notes
model_id	BAAI/bge-small-en-v1.5	Full Hugging Face model ID.
model_alias	small	The alias passed to init.
embedding_dim	384	Vector width. Determines which page_embeddings_vec_* table is used.
schema_version	10	Validated on every open.

On open, quaid verifies the requested model matches the recorded one. A mismatch errors out before any command runs. To switch models, initialize a new memory and import into it.

Build-time feature flags

Feature	Effect
embedded-model (default)	Airgapped build. Embeds BGE-small (~180 MB) directly into the binary. Non-default model selection is a warning-only no-op.
online-model	Smaller binary (~90 MB). Downloads and caches the selected model on first semantic use.

# Default (airgapped)
cargo build --release

# Online build
cargo build --release --no-default-features --features bundled,online-model

See Airgapped vs online for guidance on which channel to use, and Embedding models for the design.