Tools | termina | Documentation

The tools the Termina MCP server exposes, by area. You rarely call them by name: ask the agent a question and it chooses. Naming a tool in a prompt (“use get_cohorts for 2024”) steers it when you want a specific view. The agent receives each tool’s full description and argument schema from the server when it connects.

Every tool is read-only and runs with your Termina permissions.

Shared parameters

Most tools take the same handful of arguments.

Parameter	Meaning
`group_id`, `deal_id`	Identify the deal. Find them with `search_deals` or `get_diligence_metadata`.
`user_type`	The customer segment to read (for example `business` or `consumer`). Omit it to use the deal’s own segment; pass it only to ask about a different one.
`metric`	The measured quantity for product tools, for example `revenue`. `get_deal` lists the `user_type` and `metric` combinations a deal has.
`freq`	`M` (monthly, the default) or `Q` (quarterly).
`date_start`, `date_end`	An inclusive period range, each like `2024-01`, `2024Q1` or `2024`. Pass either or both.
`cohort_start`, `cohort_end`	The same, applied to a cohort’s start period.
`max_period`	Drops cohort periods beyond it, counted in `freq` units since the cohort began: with monthly data, `12` keeps about each cohort’s first year.

Long results are capped and marked as truncated. A request for a specific period or cohort should always carry the matching range parameters rather than rely on the cap to include it.

Discovery

Start here. A session usually opens with search_deals or get_diligence_metadata to find a group_id and deal_id, then drills into the data tools.

Tool	Description
`search_deals(query, page_size=10)`	Searches the deals you can access by free text (company, round, notes, tags) and returns compact summaries with their ids. Returns the first `page_size` matches, at most 50.
`get_deal(group_id, deal_id)`	Returns one deal’s metadata and which data exists for it: financials, talent, product metrics, unit economics, benchmark segments.
`list_companies(group_id, page_size=25)`	Lists the companies in a group, for mapping a company name to its deals. Returns the first `page_size` companies, at most 100.
`get_diligence_metadata()`	Describes every group, deal and metric type available to you. A good orientation call when you do not yet know an id.
`get_metric_definition(metric)`	Looks up Termina’s own definition of a metric, ratio or acronym (`CMGR`, `LTV/CAC`, `burn multiple`). Matching is fuzzy; up to three candidates come back best-first.

Financials

Reported figures from the company’s own statements. Margins, growth and ratios are already computed: read them from get_derived_metrics or the benchmark tools instead of working them out from the raw lines.

Tool	Description
`get_income_statement(group_id, deal_id, date_start, date_end)`	Returns the reporting frequency and unit plus one row per period and line item, with Termina’s normalized category and the company’s own label.
`get_balance_sheet(group_id, deal_id, date_start, date_end)`	Returns the balance sheet as a flattened tree: aggregate nodes carry pre-computed subtotals, leaves carry their categories, and `forecast` marks projected periods.
`get_derived_metrics(group_id, deal_id, date_start, date_end)`	Returns the financials-derived series per period: gross and operating margins, rule of 40, magic number, burn multiple, ROAS, revenue growth, and balance-sheet liquidity and leverage ratios (current, quick, cash, debt to equity).

Benchmarks

The company’s computed metrics next to Termina’s proprietary peer dataset. Call get_available_categories first when you want a peer group other than the deal’s own. Peers are anonymized: the response carries the peer count and revenue range, never names.

Tool	Description
`get_available_categories(group_id, deal_id)`	Lists the benchmark categories available for the deal’s company.
`get_benchmark_metrics(group_id, deal_id, category, user_type)`	Returns the deal view’s benchmark table: the company’s latest figures (growth, margins, capital efficiency, retention, unit economics, concentration, headcount) beside the peer distribution by percentile, plus the size of the peer cohort.
`get_benchmark_time_series(group_id, deal_id, user_type, date_start, date_end)`	Returns the monthly history of every headline benchmarked metric, in the same shape, without the peer comparison. Use it for how a figure moved over time.
`get_tradeoff_at_scale(group_id, deal_id, category)`	Returns the models of how peers trade growth against efficiency at the deal’s revenue scale.

Product and unit economics

Movement, retention, concentration and contribution computed from the company’s product data, per customer segment.

Tool	Description
`get_growth_accounting(group_id, deal_id, metric, user_type, freq, date_start, date_end)`	Decomposes each period’s amount into new, retained, resurrected, expansion, contraction and churned, with gross retention, quick ratio, net churn and growth rates. Explains why a figure moved.
`get_user_accounting(group_id, deal_id, metric, user_type, freq, date_start, date_end)`	The same decomposition on user counts rather than revenue, to separate logo churn from dollar churn.
`get_cohorts(group_id, deal_id, metric, user_type, freq, cohort_start, cohort_end, max_period)`	Returns the retention curve: one row per cohort and period with active users, amount, logo and amount retention, and cumulative LTV per user.
`get_customer_cohorts(group_id, deal_id, metric, user_type, freq, cohort_start, cohort_end, max_period)`	Returns per-customer rows for the top 25 customers by ACV and lifetime value, to see which customers drove a cohort’s movement. Customers are opaque ids.
`get_concentration(group_id, deal_id, metric, user_type, freq)`	Returns the share held by the top 5, 10 and 20 customers and the top 10% and 20%, with mean, median, percentiles and Gini coefficient per period, plus the largest customers and the cumulative distribution for the latest period.
`get_contribution(group_id, deal_id, user_type, freq, cohort_start, cohort_end, max_period)`	Returns unit economics per cohort and period: CAC, LTV, gross-margin LTV, LTV/CAC and contribution, with `method` separating actual rows from projected ones.
`get_contribution_summary(group_id, deal_id, user_type, freq)`	Returns the deal view’s unit economics summary: headline LTV, CAC, modeled payback and margins per cohort, with lifetime and rolling variants.

Talent

Tool	Description
`get_executive_profiles(group_id, deal_id)`	Returns the leadership roster: each executive’s position history with titles, dates and which roles were at this company.
`get_team_growth_accounting(group_id, deal_id, date_start, date_end)`	Returns headcount by function and quarter, decomposed into new, retained and resurrected, for questions like which functions the company is hiring into.
`get_seniority_count(group_id, deal_id)`	Returns the current headcount by seniority level, as a single snapshot.

Documents

Two different documents: what Termina’s analyst wrote about the deal, and what the company put in its own pitch deck. The write-up is the primary source for narrative; the deck is secondary and its figures are cited by page.

Tool	Description
`get_deal_document(group_id, deal_id)`	Returns the analyst-authored deal document (often called the memo): narrative, key findings with their signal, embedded charts and tables, reader comments and chart annotations.
`get_investor_deck_data(group_id, deal_id)`	Returns the parsed investor deck the company uploaded: per-page text, chart and table data series, and context flags, with 1-indexed page numbers for citation. Long page text and long series are clipped with a note. Empty when no deck has been parsed.

Macro

Market-wide signals that are not tied to a deal. Call list_market_segmentations first to learn the valid segmentation names and slugs.

Tool	Description
`list_market_segmentations()`	Lists the available market segmentations, such as geography and sector, with their slugs.
`get_market_gauges(segmentation, slug)`	Returns the market “weather” gauge per date: a score from -3 to +3 for how fundraising compares with what conditions would predict, where 0 is normal and negative is colder. Omit `slug` for every market in the segmentation.
`get_market_context(segmentation, slug)`	Returns rolling estimates of funding activity per date and stage: annualized amount raised and round count, years between raises, and average raise.