CLI Reference

Global Flags

Available for all commands:

--no-color – Disable colored output (also respects NO_COLOR env var)
-v, --verbose – Verbose output
-q, --quiet – Suppress non-error output

`srake ingest`

Ingest SRA metadata from NCBI or local archives.

srake ingest [flags]

Source flags:

Flag	Description
`--auto`	Auto-select the best file from NCBI
`--daily`	Ingest the latest daily update
`--monthly`	Ingest the latest monthly dataset
`--file <path>`	Ingest a local or remote file
`--list`	List available files without ingesting

Filter flags:

Flag	Description
`--taxon-ids <ids>`	Comma-separated NCBI taxonomy IDs to include
`--exclude-taxon-ids <ids>`	Taxonomy IDs to exclude
`--organisms <names>`	Organism names to include
`--platforms <names>`	Sequencing platforms (e.g. ILLUMINA, PACBIO_SMRT)
`--strategies <names>`	Library strategies (e.g. RNA-Seq, WGS)
`--date-from <date>`	Include records from this date (YYYY-MM-DD)
`--date-to <date>`	Include records until this date
`--min-reads <n>`	Minimum read count
`--max-reads <n>`	Maximum read count
`--min-bases <n>`	Minimum base count
`--max-bases <n>`	Maximum base count
`--stats-only`	Preview filter results without inserting

Other flags:

Flag	Description
`--db <path>`	Database path
`--force`	Force re-ingestion
`--no-progress`	Disable progress bar

# Examples
srake ingest --auto
srake ingest --file archive.tar.gz --taxon-ids 9606 --platforms ILLUMINA
srake ingest --auto --stats-only  # preview what would be imported

`srake search`

Search SRA metadata using full-text, vector, or hybrid search.

srake search <query> [flags]

Filter flags:

Flag	Description
`--organism <name>`	Filter by organism
`--platform <name>`	Filter by platform
`--library-strategy <name>`	Filter by library strategy
`--library-source <name>`	Filter by library source
`--library-selection <name>`	Filter by library selection
`--library-layout <name>`	Filter by library layout
`--study-type <name>`	Filter by study type
`--instrument-model <name>`	Filter by instrument model
`--date-from <date>`	Date range start
`--date-to <date>`	Date range end
`--spots-min <n>`	Minimum spots (reads)
`--spots-max <n>`	Maximum spots
`--bases-min <n>`	Minimum bases
`--bases-max <n>`	Maximum bases

Output flags:

Flag	Description
`--limit <n>`	Max results (default: 100)
`--offset <n>`	Skip N results
`--format <type>`	Output format: table, json, csv, tsv, accession
`--output <file>`	Write results to file
`--no-header`	Omit table header
`--fields <list>`	Comma-separated field list

Search mode flags:

Flag	Description
`--search-mode <mode>`	Search mode: auto, text, vector, hybrid, database
`--fuzzy`	Enable fuzzy matching
`--exact`	Require exact matches
`--advanced`	Enable advanced query syntax
`--similarity-threshold <f>`	Vector similarity threshold (0.0-1.0, default: 0.5)
`--min-score <f>`	Minimum BM25 score
`--show-confidence`	Show confidence scores
`--hybrid-weight <f>`	Hybrid weight (0.0=text, 1.0=vector, default: 0.7)
`--facets`	Include facet counts
`--stats`	Show search statistics

# Examples
srake search "homo sapiens"
srake search "cancer" --organism "homo sapiens" --format json
srake search "tumor expression" --search-mode vector --show-confidence
srake search "RNA-Seq" --format accession --output accessions.txt

`srake index`

Build and manage the search index.

srake index [flags]

Flag	Description
`--build`	Build the search index
`--rebuild`	Rebuild from scratch
`--verify`	Verify index integrity
`--stats`	Show index statistics
`--resume`	Resume interrupted build
`--batch-size <n>`	Documents per batch (default: 500)
`--workers <n>`	Parallel workers (0 = auto)
`--path <dir>`	Index directory
`--backend <type>`	Backend: tiered (default), bleve
`--with-embeddings`	Include vector embeddings
`--embedding-model <name>`	Embedding model name
`--progress`	Show progress bar

# Examples
srake index --build
srake index --build --with-embeddings --progress
srake index --rebuild --batch-size 1000
srake index --stats

`srake server`

Start the REST API server.

srake server [flags]

Flag	Description
`-p, --port <n>`	Port (default: 8080)
`--host <addr>`	Host to bind (default: 0.0.0.0)
`--enable-cors`	Enable CORS (default: true)
`--db <path>`	Database path
`--index <path>`	Index path

# Examples
srake server --port 8080
srake server --port 3000 --host localhost
SRAKE_DB_PATH=/data/srake.db srake server

See API Reference for endpoint documentation.

`srake mcp`

Start an MCP (Model Context Protocol) server for AI assistants.

srake mcp [flags]

Flag	Description
`--db <path>`	Database path
`--index <path>`	Index path
`--transport <mode>`	Transport mode: `stdio` (default) or `http`
`--host <addr>`	HTTP server host (default: localhost, only with `--transport http`)
`--port <n>`	HTTP server port (default: 8081, only with `--transport http`)

Stdio mode (default): communicates over stdin/stdout using JSON-RPC 2.0. All diagnostic output goes to stderr.

HTTP mode: starts a Streamable HTTP server. MCP clients connect over the network using the Streamable HTTP transport spec.

Tools: search_sra, get_metadata, find_similar, export_results

Resources: sra://stats, sra://search/recent, sra://studies/{accession}

Prompts: biomedical_search, sample_selection

# Claude Desktop configuration (~/.claude/claude_desktop_config.json):
{
  "mcpServers": {
    "srake": {
      "command": "srake",
      "args": ["mcp"]
    }
  }
}

# With custom database
srake mcp --db /path/to/srake.db

# HTTP transport for remote access
srake mcp --transport http --port 8081

# HTTP transport on all interfaces
srake mcp --transport http --host 0.0.0.0 --port 9090

`srake metadata`

Retrieve metadata for specific accessions.

srake metadata <accession> [accessions...] [flags]

Flag	Description
`-f, --format <type>`	Output format: table, json, yaml
`--fields <list>`	Comma-separated field list
`--expand`	Expand nested structures

Supports SRP/DRP/ERP (study), SRX/DRX/ERX (experiment), SRS/DRS/ERS (sample), and SRR/DRR/ERR (run) accessions.

# Examples
srake metadata SRX123456
srake metadata SRP000001 --format json

`srake db`

Database management commands.

`srake db info`

Show database statistics.

srake db info

`srake db stats`

Manage pre-computed statistics.

srake db stats --show
srake db stats --rebuild

`srake db export`

Export the database to SRAmetadb.sqlite format.

srake db export [flags]

Flag	Description
`-o, --output <file>`	Output path (default: SRAmetadb.sqlite)
`--db <path>`	Source database path
`--fts-version <n>`	FTS version: 3 or 5 (default: 5)
`--batch-size <n>`	Batch size (default: 10000)
`--progress`	Show progress (default: true)
`--compress`	Gzip compress output
`-f, --force`	Overwrite existing file

# Examples
srake db export -o SRAmetadb.sqlite
srake db export -o legacy.sqlite --fts-version 3
srake db export -o SRAmetadb.sqlite.gz --compress

See SRAmetadb Export for details.

`srake models`

Manage embedding models for vector search.

`srake models list`

List installed models.

`srake models download <model-id>`

Download a model. Flag: --variant (quantized, fp16, full).

`srake models test <model-id> <text>`

Test a model with sample text.

# Examples
srake models list
srake models download Xenova/SapBERT-from-PubMedBERT-fulltext --variant quantized
srake models test Xenova/SapBERT-from-PubMedBERT-fulltext "breast cancer"

Environment Variables

Variable	Description
`SRAKE_DB_PATH`	Database path
`SRAKE_INDEX_PATH`	Search index path
`SRAKE_MODELS_PATH`	Models directory
`SRAKE_EMBEDDINGS_PATH`	Embeddings directory
`SRAKE_MODEL_VARIANT`	Model variant: full, quantized, fp16
`SRAKE_SEARCH_BACKEND`	Search backend: tiered, bleve, sqlite
`SRAKE_CONFIG`	Config file path
`NO_COLOR`	Disable colored output
`XDG_CONFIG_HOME`	XDG config directory
`XDG_DATA_HOME`	XDG data directory
`XDG_CACHE_HOME`	XDG cache directory
`XDG_STATE_HOME`	XDG state directory

See Configuration for full details.

Configuration