Specter

Get a company's IPO events by Specter organization ID: listing date, exchange and ticker, share price, and the valuation and capital raised at listing. Supports pagination. Render as a table: Date | Exchange | Ticker | Share Price | Valuation | Raised. Events are ordered most recent first. A company with no public listing returns an empty `ipos`. Omit null fields.

get_company_ipos

Get company intelligence by Specter organization ID: department headcount, investor signals, key people, founders, highlights, and the latest revenue and valuation estimates. Headcount: from `headcount_by_department`, show only the most recent month as a table (department, count, % of total). Highlight engineering % as an R&D investment proxy. Investor signals: render as a table (Date | Source | Score | Type). Scores 1-10. Include description when present. Team: render founders and key people with background highlights. Highlights: translate to plain English bullets. Financial position: `latest_revenue` and `latest_valuation` are already here — no follow-up call needed. Report revenue with its period and valuation with its date, and say plainly when either is null; an absent figure is not zero.

get_company_intelligence

Get investor-interest signals for a company by Specter organization ID: the funding, growth, and market-activity signals that indicate investor attention on that company. Supports date-range filtering (`date_from` / `date_to`, on the signal date) and pagination. Render as a table: Date | Score | Type | Summary. Scores are 1-10. Signals are ordered newest first. `signal_investors` names the investors behind a signal — offer `get_investor_profile` for any of them. Omit null fields. `get_company_intelligence` returns a short inline preview of the same signals; use this tool when the full or date-filtered feed is wanted.

get_company_investor_interest_signals

Get a company's latest revenue by Specter organization ID: the most recent reported revenue figure (with its date and metric) and the ML-estimated revenue with its growth trend. Report `latest_revenue_reported_usd` (with date and `revenue_type`) as the reported figure and `est_revenue_usd` as the Specter estimate. Format money compactly (e.g. $120M). Growth percentages: 4.05 means +4.05% (can be negative). Omit null fields.

get_company_latest_revenue

Get a company's latest valuation by Specter organization ID: the most recent reported post-money valuation (with its date) and the ML-estimated valuation with its growth trend. Report `latest_valuation_reported_usd` (with date) as the reported figure and `est_valuation_usd` as the Specter estimate. Format money compactly (e.g. $1.2B). Growth percentages: 4.05 means +4.05% (can be negative). Omit null fields.

get_company_latest_valuation

Get news signals attributed to a company by Specter organization ID: funding, revenue, traction, profit, and deal updates surfaced from news coverage. Supports date-range filtering (`date_from` / `date_to`, on the article publication date) and pagination. Render as a table: Date | Title (hyperlinked to `article_url`) | Updates | Score | What's New. Scores are 1-10 (>= 4 is major news). Signals are ordered most-recent first. When `total` exceeds the number of returned signals, offer to fetch the next page.

get_company_news_signals

Get a detailed company profile by Specter organization ID. First-time presentation: render as a mini investment memo. Memo skeleton (skip empty sections): 1. Snapshot — name (linked to `website`), HQ, founded, employees, growth_stage, total_funding, monthly web visits. 2. What They Do — short_description, typical_customer_profile, traction. 3. Industry & Tech Verticals — see below. 4. Funding History — table (call `get_company_funding_rounds`; add `get_company_acquisitions` / `get_company_ipos` when the company has exited). 5. Investor Signals — table (call `get_company_intelligence`). 6. Team — founders + key people (call `get_company_intelligence`). 7. Headcount — most recent month (call `get_company_intelligence`). 8. Reported Clients. Hyperlink `name` to `https://app.tryspecter.com/signals/company/feed/{external_company_id}`. Show `linkedin_url` as a clickable link. Omit null fields. Render `traction` as a bullet list, not the raw comma-separated string. Industry & Tech Verticals (proprietary dual-lens classification, complementary — show both when present): - `industry`: 3-level economic-activity hierarchy (21 L1 → L2 → L3). Render with ` > ` separator. - `tech_verticals`: 2-level tech-domain hierarchy (33 verticals → 300+ sub-verticals). Render with ` > ` separator.

get_company_profile

The reverse of `get_list_items`: for a batch of records, which Specter lists they are in and what the team's connected CRM holds on them (CRM lists and per-entry status, record owner, lifecycle stage, linked deals, last/next activity). Reach for it whenever the answer depends on what the team already has on a record rather than on what Specter knows about it — annotating a search page or list with the CRM list each record sits in, briefing on several CRM lists at once, or telling saved records apart from new ones. Pass the external ids another tool returned, matching `product_type`. Every requested id comes back, in the order asked: an empty `specter_lists` means the record is in none the caller can access, and an empty `crm_records` means no connected CRM holds it. `crm_supported` is false for investors, talent and stratintel, whose `crm_records` are empty for want of a CRM counterpart rather than for want of a match — say so rather than reporting them as absent from the CRM. CRM values are the team's own, mirrored at `synced_at` and shown as the CRM names them; render them verbatim. Only the pulled fields listed above exist — a CRM's custom properties are not mirrored, so never imply a record has none of something this tool cannot see. Render as one row per record, naming the Specter lists and, per CRM, the lists with their status.

get_entity_memberships

Get funding rounds a given investor participated in by Specter investor ID, with the funded company and co-investors. Supports date-range filtering (`date_from` / `date_to`, on the announcement date) and pagination. Render as a table: Date | Company | Round | Raised | Lead. Rounds are ordered newest first. When `total` exceeds the returned count, offer to fetch the next page. Omit null fields.

get_investor_funding_rounds

Get investor-interest signals sourced from a given investor by Specter investor ID: the funding, growth, and market-activity signals attributed to that investor. Supports date-range filtering (`date_from` / `date_to`, on the signal date) and pagination. Render as a table: Date | Entity | Score | Type | Summary. Scores are 1-10. Signals are ordered newest first. `entity_id` links the company / person the signal concerns — offer `get_company_profile` / `get_person_profile`. Omit null fields.

get_investor_interest_signals

Get a detailed investor profile by Specter investor ID: firmographics, investment activity (counts, stages, industries, tech verticals), and portfolio company IDs. Render as a mini profile: name, type, HQ, founded, then activity (investments / lead / exits / funds) and focus (stages, industries, tech verticals). Show `linkedin_url` / `crunchbase_url` as links. When `company_id` or `person_id` is present the investor is itself an entity — offer `get_company_profile` / `get_person_profile`. Omit null fields.

get_investor_profile

Get items in a specific list by list ID. Use `get_lists` first to resolve a list name to an ID. Show list name, item count, and list ID at the top. Render items as a concise table: name (hyperlinked), social links, key metric (stage, funding, employees). Hyperlink companies to `https://app.tryspecter.com/signals/company/feed/{external_company_id}` and people to `https://app.tryspecter.com/signals/people/table/{external_person_id}`. Offer to pull full details on any result. For company items, render `industry` and `tech_verticals` with ` > ` separator when present. Investor items: render name, type, HQ, #investments and #exits; investors are not hyperlinked.

get_list_items

List accessible lists for the current user. Supports optional name filtering and pagination. Render as a concise table: list name, item count, list ID. Offer to pull items via `get_list_items`.

get_lists

Get a personal or professional email address for a person by external person ID.

get_person_email

Get a detailed person profile by external person ID, including role history and education. First-time presentation: {full_name} · LinkedIn {tagline} · Current: {current_position.title} at {current_position.company_name} Then: career timeline (most recent first) → education → highlights (plain English). Hyperlink `name` to `https://app.tryspecter.com/signals/people/table/{external_person_id}`. Show `linkedin_url` as clickable link. Surface `top_university` from education when true. Highlight: domain expertise, founder-market fit, prior exits, track record, years of experience.

get_person_profile

Get the talent signals for one person by Specter person ID: the career moves Specter tracked for them — founding a company, going stealth, coming out of stealth, executive hires, promotions. Supports date-range filtering (`date_from` / `date_to`, on the signal date) and pagination. Render as a table: Date | Signal | From → To | Score. Scores are 1-10. Signals are ordered newest first. `new_position_company_external_id` is a Specter company ID — offer `get_company_profile` for it. Surface `out_of_stealth_advantage` when present as how many days early the move was caught. Append (Est.) to industries and tech verticals when `is_industry_estimated` is true. When `total` exceeds the number of returned signals, offer to fetch the next page. Omit null fields.

get_person_talent_signals

List results returned by a specific saved search ID. Use `get_saved_searches` first to resolve a search name to an ID. The product — company, people, talent signals, investor interest signals, investors (firms), funding rounds, acquisitions, IPOs — is inferred from the saved search; you do not pass it. For a people search, `signal_mode='new_talent_signals'` returns TalentSignalItem entries, not PeopleSearchItem. Show search name, item count and search ID at the top. Render `industry` / `industries` and `tech_verticals` with a ` > ` separator, noting the industry is estimated when `is_industry_estimated` is true. Signal scores are 1-10. Default — concise table: name (hyperlinked), social links, key metric. Hyperlink companies to `https://app.tryspecter.com/signals/company/feed/{external_company_id}`, people to `https://app.tryspecter.com/signals/people/feed/{external_person_id}`. Talent signals (people movements: stealth founders, exec departures, new ventures): Date | Person | Title | Company | Score | Signal Type | Signal Summary | Industry | Tech Vertical. Always show `signal_summary` (background + why notable) as a brief narrative below the row; a stealth company explains an estimated industry. Offer `get_person_profile` / `get_person_email` for follow-up. Investor interest signals (investor activity patterns): Date | Name | Score | Signal Type | Signal Summary | Industry | Tech Vertical. Investors (firms, e.g. VC funds): Name | Type | HQ | #Investments | #Exits — not hyperlinked. Funding rounds: Date | Company | Round | Raised (USD) | #Investors. Acquisitions: Date | Acquirer | Acquiree | Type | Price (USD), saying undisclosed rather than leaving a blank. IPOs: Date | Company | Exchange:Ticker (`stock_exchange_symbol`:`stock_symbol`) | Share Price (USD) | Raised (USD) | Valuation (USD). Hyperlink each company to `https://app.tryspecter.com/signals/company/feed/{external_company_id}` via `company_id`, or `acquirer_company_id` / `acquiree_company_id` for acquisitions.

get_saved_search_items

List saved searches accessible to the current user, for any product: company, people, talent signals, investor interest signals, investors (firms), and the transaction feeds — funding rounds, acquisitions and IPOs. Supports optional name filtering and pagination. Render as a concise table: search name, item count, search ID. The count fields are named per product (e.g. `full_companies_count`, `full_funding_rounds_count`, `full_acquisitions_count`, `full_ipos_count`). Offer to pull results via `get_saved_search_items`. `last_viewed_at` is when the current user last opened each search. For "what is new since I last checked", do not read it and do not ask for a date — call `get_saved_search_items` with `since_last_viewed=true`, which resolves the same stored moment itself. To link a search, open its product feed with BOTH `queryId=<query_id>` and `searchId=<search_id>`.

get_saved_searches

Return the Specter MCP usage instructions and presentation guidelines.

get_usage_instructions

Run a Specter search from a natural-language DESCRIPTION of the records you want (e.g. 'climate-focused VC investors', 'high-growth fintech companies in Germany') and — when you know the kind of record you want — set `product`. Omit it to let Specter infer the product from the query. This is for DISCOVERY, never for looking up an entity you can already name — 'find Stripe' / 'the person Jane Doe' belongs in `find_company` / `find_person`, which resolve a known name or domain directly. `product` is one of: company, people, investors, talent_signals, interest_signals, funding_rounds, acquisitions, ipos. Use `investors` for the investor firms themselves (e.g. VC funds), not their portfolio companies. Returns a compact list per product — the same fields the saved-search tools return, not the full record. Drill into a single result via the per-record tools (`get_company_profile`, `get_person_profile`, …) for the complete object. Also returns `query_id` / `agent_query_id`, a `search_url` to reopen the whole search in the app, and the `product` actually searched. To order the results, pass `sort` — a list of `{field, direction}` (`direction` is `asc` or `desc`, default `desc`). Which fields a product accepts is product-specific and runs well beyond the obvious few; an unknown one is rejected with that product's full list, so an uncertain guess costs a call rather than the ranking. People searches can't be sorted. When rendering results, hyperlink each item's name to its Specter page: `company` (field `id`), `funding_rounds`/`ipos` (field `company_id`) and `acquisitions` (field `acquiree_company_id`) link to a company feed at `https://app.tryspecter.com/signals/company/feed/{external_company_id}`; `people` (field `id`) and `talent_signals` (field `specter_person_id`) link to `https://app.tryspecter.com/signals/people/feed/{external_person_id}`. `investors` and `interest_signals` are not hyperlinked.

search

Update a list: rename it, change whether it is shared with the API, and/or change its membership. Works for any list product (the product is resolved from the list ID, so item ids must match that product). Use `get_lists` to resolve a list name to its ID and `get_list_items` to see what is in it. Membership can be changed on any list you can access; renaming and `is_shared_with_api` are limited to lists you own or administer. Pass at least one of `name` / `is_shared_with_api` / `add_item_ids` / `remove_item_ids`. Item ids are the external ids returned by other tools: company → `external_company_id`, people → `external_person_id` (per_...), investors → investor id, talent → talent id, stratintel → strategic id. Membership changes are idempotent and reported per direction: items already in the list count towards `already_present`, requested removals that were not members towards `not_present` — neither is an error. Removals are applied after additions. An id that is malformed for the list's product, or that names no entity Specter knows, fails the whole call and nothing is written — resolve ids with a find/search tool first. Fails too if you already own another list of the same product with the requested name.

update_list