# stocks_get_peers Comparable companies for one company, with fundamentals side by side. | Field | Value | | --- | --- | | Category | Stocks (`stocks`) | | Exposure | mixed | | MCP tool | `stocks_get_peers` | | Documentation | https://www.agentdatasets.com/docs/tools/stocks_get_peers | ## Public-key response SEC-derived fundamental and statement metrics are available to public keys. Peer comparisons that request price-derived market cap, P/E, P/S, P/B, EV/EBITDA, or FCF yield are refused; those metrics require an internal key. ## Description Comparable companies for one company, with fundamentals side by side. Use this right after resolving a company, when the question is "who does this trade against?" or "how does it compare to its peers?" — it returns the comparison set and the metrics in one bounded call. Use `stocks_screen` instead when you already know the criteria and want to rank or page the whole cross-section, and `stocks_get_financial_metrics` when you only need one company's own multi-period detail. Peers are active-listed companies sharing the subject's classification in the platform's own SIC-derived taxonomy (SEC SIC codes mapped by us; no licensed vendor taxonomy is involved). `data.basis` says which cut was used: `sic-major-group` — same 2-digit SEC SIC major group inside the same sector, the tighter and preferred set when the subject has a SIC code — or `sector`, the fallback when the subject has no SIC code or its major group contains no other listed company. `data.sector` and `data.sic_major_group` echo the classification used. Ordering is nearest fundamental scale: ascending absolute log-ratio of trailing-annual `revenue` against the subject, so a peer twice the size ranks the same distance away as one half the size. Companies with no comparable revenue value sort last, alphabetically. `data.subject` is the subject's own row (it is never listed among its peers) and `data.peers` holds the comparison rows. Every row carries `ticker`, `cik`, `company_name`, `exchange`, `sector`, `sic_code`, and `metrics`: exact-decimal values with unit, currency, fiscal frame, `as_of`, and source. `data.peer_universe_size` is how many companies the basis matched — larger than the returned list means you are seeing the closest slice. `meta.as_of` is when the metrics snapshot last rebuilt. A subject we have not classified (no SIC, or one outside our mapping) returns `basis: null`, `peers: []`, and a `reason` saying so — a normal success, not an error. Results are bounded: default 10 peers, maximum 20, and there is no paging (`pagination` is null) — this is a nearest-N answer, so use `stocks_screen` to enumerate a full cross-section. Metrics are bounded at 10 per row, counted on what is served: `revenue` is always one of them, so name at most 10 metrics including `revenue`, or at most 9 that are not — 10 non-revenue names is one too many and is rejected as `bad_parameter`. Errors carry a machine `code`: `unknown_entity` for a symbol or CIK we hold no company for, `bad_parameter` for an unknown metric name or an out-of-range bound, and `out_of_license` when a public request names a price-derived metric (`market_cap`, `pe_ttm`, `ps_ttm`, `pb`, `ev_ebitda`, `fcf_yield`) — the same gate `stocks_screen` applies. Example: `stocks_get_peers(ticker_or_cik="AAPL", limit=10, metrics=["revenue", "net_margin", "roe"])`. ## Parameters Input schema advertised to MCP clients and mirrored by the REST query string. | Parameter | In | Type | Required | Default | Constraints | Description | | --- | --- | --- | --- | --- | --- | --- | | ticker_or_cik | input | `string` | yes | — | — | The subject company: a ticker symbol (AAPL, BRK.B) or a CIK (320193 or 0000320193). | | limit | input | `integer` | no | 10 | min 1, max 20 | Peers to return, nearest in size first (1-20); default 10. Out-of-range values are rejected. | | metrics | input | `array \| null` | no | null | max items 10 | Snapshot metric names to compare on, from the `stocks_screen` vocabulary. Omit for the default set (revenue, revenue_growth_yoy, net_margin, roe). `revenue` is always included because the ranking is measured on it, so every row carries at most 10 metrics: name up to 10 including `revenue`, or up to 9 that are not `revenue`. | ## Provenance - `sec_edgar_companyfacts` — U.S. Securities and Exchange Commission (EDGAR), XBRL company-facts financial-statement data (per-company API and the nightly bulk companyfacts.zip) (U.S. Government work, public domain (17 U.S.C. § 105)). See https://www.agentdatasets.com/docs/attribution.md - `tiingo` — provider-licensed source; not served to public keys Dataset registry: `stocks_daily_prices`, `stocks_fundamentals`.