### NCBI Entrez ESearch (`ncbi-esearch`) Runs a query term against a chosen Entrez database and returns the list of matching record UIDs, with optional pagination, sort key, single-field restriction, and date filtering on a modification, publication, or Entrez date axis. #### API Reference

Source NCBI database to query (e.g. 'protein', 'nuccore', 'gene', 'pubmed', 'taxonomy', 'structure'). Available options: `protein`, `nuccore`, `nucleotide`, `gene`, `pubmed`, `pmc`, `taxonomy`, `structure`, `snp`, `clinvar`, `omim`, `biosample`, `bioproject`, `sra`, `assembly`, `ipg`, `mesh`, `genome`, `dbvar`, `gds`, `geoprofiles`, `medgen`, `proteinclusters`, `protfam`, `pccompound`, `pcsubstance`, `pcassay` NCBI search query. Max IDs returned (NCBI `retmax`). 0-indexed offset of the first hit (NCBI `retstart`). Sort key (db-dependent — e.g. 'relevance' / 'pub\_date' / 'most\_recent' on pubmed). Restrict the search term to a single index field (db-dependent — e.g. 'title' / 'author' on pubmed). Date axis for mindate/maxdate/reldate (modification / publication / Entrez). Lower date bound, `YYYY/MM/DD` (also `YYYY/MM` and `YYYY`); requires datetype. Upper date bound; requires datetype. Restrict to records dated within the last N days; requires datetype.

Source Verbosity level (0=quiet, 1=info, 2=debug, 3=raw subprocess stderr). `True` is coerced to `1` and `False` to `0`. Device to run the tool on. Maximum execution time in seconds. `None` waits indefinitely. Random seed. When set, tools run reproducibly up to small GPU float noise (see `BaseToolOutput.approx_equal`), and the seed participates in cache keys. When None, cacheable seed-sensitive tools skip cache until seeded. Optional NCBI API key (lifts rate limit from 3 to 10 requests/second). Defaults to the `NCBI_API_KEY` environment variable; an explicit value passed to the config overrides the env var. Optional contact email. NCBI usage policy requires both `tool` and `email` for traceability. Defaults to the `NCBI_EMAIL` environment variable; an explicit value passed to the config overrides the env var.

Source List of NCBI IDs matching the search query.

#### Applications Use this as the entry point of an Entrez retrieval pipeline: resolve a gene symbol and organism to candidate protein or nucleotide UIDs, page through a large hit set with `retstart` and `max_results`, or restrict a literature query by date before downstream processing. The returned UIDs feed directly into [`ncbi-esummary`](https://bio-pro.mintlify.app/tools/database-retrieval/ncbi) for metadata screening and [`ncbi-efetch`](https://bio-pro.mintlify.app/tools/database-retrieval/ncbi) for sequence retrieval, and a resolved accession pairs naturally with the [UniProt](https://bio-pro.mintlify.app/tools/database-retrieval/uniprot) and [sequence-fetch](https://bio-pro.mintlify.app/tools/database-retrieval/sequence-fetch) tools. #### Usage Tips * **The returned IDs are Entrez UIDs, not always accessions.** Depending on the database they may be numeric GI numbers (GenInfo Identifiers). Resolve them through `ncbi-esummary` or `ncbi-efetch` to obtain accession-bearing records. * **Date bounds need a date axis.** Setting `mindate`, `maxdate`, or `reldate` without `datetype` is rejected, because NCBI silently ignores date filters that lack an axis. * **`sort` and `field` are database-specific.** A key valid on `pubmed` may be invalid on `protein`. Consult the Entrez help for the database being queried.

### NCBI Entrez ESummary (`ncbi-esummary`) Retrieves the document summary for a UID or accession from a chosen Entrez database and returns the summary as a database-specific mapping alongside the sanitized request URL. #### API Reference

Source NCBI database to query (e.g. 'protein', 'nuccore', 'gene', 'pubmed', 'taxonomy', 'structure'). See `NCBIDatabase` for the full set of supported databases. Available options: `protein`, `nuccore`, `nucleotide`, `gene`, `pubmed`, `pmc`, `taxonomy`, `structure`, `snp`, `clinvar`, `omim`, `biosample`, `bioproject`, `sra`, `assembly`, `ipg`, `mesh`, `genome`, `dbvar`, `gds`, `geoprofiles`, `medgen`, `proteinclusters`, `protfam`, `pccompound`, `pcsubstance`, `pcassay` Accession or NCBI ID to summarize (e.g. 'NP\_000537.3', '7157').

Source Verbosity level (0=quiet, 1=info, 2=debug, 3=raw subprocess stderr). `True` is coerced to `1` and `False` to `0`. Device to run the tool on. Maximum execution time in seconds. `None` waits indefinitely. Random seed. When set, tools run reproducibly up to small GPU float noise (see `BaseToolOutput.approx_equal`), and the seed participates in cache keys. When None, cacheable seed-sensitive tools skip cache until seeded. Optional NCBI API key (lifts rate limit from 3 to 10 requests/second). Defaults to the `NCBI_API_KEY` environment variable; an explicit value passed to the config overrides the env var. Optional contact email. NCBI usage policy requires both `tool` and `email` for traceability. Defaults to the `NCBI_EMAIL` environment variable; an explicit value passed to the config overrides the env var.

Source Record summary data returned by esummary. Sanitized URL used for the request.

#### Applications Use this to screen candidates cheaply before fetching full records: take the UIDs from [`ncbi-esearch`](https://bio-pro.mintlify.app/tools/database-retrieval/ncbi), inspect titles, lengths, or organism fields in the summary, and select the canonical record before paying the cost of a sequence download with [`ncbi-efetch`](https://bio-pro.mintlify.app/tools/database-retrieval/ncbi). A `gene`-database summary also bridges organism-level data to protein-centric records in the [UniProt](https://bio-pro.mintlify.app/tools/database-retrieval/uniprot) tool by resolving the canonical gene symbol. #### Usage Tips * **The summary shape depends on the database.** A `gene` summary nests fields under the UID key while `protein` and `nuccore` summaries expose record fields directly. Read the structure for the database queried rather than assuming a fixed schema. * **Multiple identifiers can be summarized in one call.** Passing a comma-joined list of UIDs returns one entry per UID, which is the efficient way to screen a full `ncbi-esearch` hit set. * **A missing record raises rather than returning empty.** An unresolved database-and-identifier pair raises an error, so guard identifiers that may be obsolete or suppressed.

### NCBI Entrez EFetch (`ncbi-efetch`) Fetches full sequence records by UID or accession from the protein, nuccore, or nucleotide databases, returning parsed FASTA records and the sanitized request URL, with optional subsequence and strand selection. #### API Reference

Source Sequence database to query. Available options: `protein`, `nuccore`, `nucleotide` Accession or NCBI ID to fetch (e.g. 'NP\_000537.3'). NCBI rettype. 'fasta\_cds\_na' is nuccore-only. Available options: `fasta`, `fasta_cds_na` Subsequence start (1-indexed, inclusive). Subsequence stop (1-indexed, inclusive). Strand for nucleotide retrieval.

Source Verbosity level (0=quiet, 1=info, 2=debug, 3=raw subprocess stderr). `True` is coerced to `1` and `False` to `0`. Device to run the tool on. Maximum execution time in seconds. `None` waits indefinitely. Random seed. When set, tools run reproducibly up to small GPU float noise (see `BaseToolOutput.approx_equal`), and the seed participates in cache keys. When None, cacheable seed-sensitive tools skip cache until seeded. Optional NCBI API key (lifts rate limit from 3 to 10 requests/second). Defaults to the `NCBI_API_KEY` environment variable; an explicit value passed to the config overrides the env var. Optional contact email. NCBI usage policy requires both `tool` and `email` for traceability. Defaults to the `NCBI_EMAIL` environment variable; an explicit value passed to the config overrides the env var.

Source Parsed FASTA records from efetch. FASTA header line (without >). Sequence string with whitespace stripped. Best-effort accession extracted from header. Sanitized URL used for the request.

#### Applications Use this to pull reference sequences into a design or analysis pipeline: retrieve a wild-type protein before sequence design, fetch a coding DNA sequence with `return_format="fasta_cds_na"` for codon-usage analysis, or extract a defined genomic region for regulatory-element work. It is the final stage of the canonical Entrez chain, consuming UIDs produced by [`ncbi-esearch`](https://bio-pro.mintlify.app/tools/database-retrieval/ncbi) and screened with [`ncbi-esummary`](https://bio-pro.mintlify.app/tools/database-retrieval/ncbi), and complements the [UniProt](https://bio-pro.mintlify.app/tools/database-retrieval/uniprot) and [sequence-fetch](https://bio-pro.mintlify.app/tools/database-retrieval/sequence-fetch) tools for cross-source retrieval. #### Usage Tips * **Restricted to sequence databases.** Only `protein`, `nuccore`, and `nucleotide` return sequence bodies. Metadata databases require `ncbi-esummary` instead. * **`fasta_cds_na` requires a nucleotide database.** This return format extracts coding DNA and is rejected for `db="protein"`, since CDS extraction has no meaning on a protein record. * **Subsequence coordinates are 1-indexed and inclusive on both ends, to match biological residue selection conventions.** Position 1 is the first residue, and `seq_stop` is included in the returned span. * **Strand `"-"` returns the reverse complement.** Antisense retrieval applies to nucleotide databases. It returns the reverse complement of the requested region.