Dealroom Company Intelligence Scraper

Dealroom Company Intelligence Scraper collects Dealroom market-map, company lookup, funding signal, and newly founded startup records into normalized JSON. It supports Dealroom list URLs, company URLs, market-map ids, company name lookup, live signal feeds, sorting, capped runs, and optional connector export.

Why use it

• Source startup and scaleup lists from public Dealroom market maps.
• Resolve company names to Dealroom profile URLs and metadata availability flags.
• Track public funding and market signals for research workflows.
• Export clean records to CRM, warehouse, spreadsheet, or MCP-connected apps.
• Run predictable samples with one finite default cap.

Input

Field	Type	Default	Description
mode	string	marketmaps	Collection mode: market-map search, market-map companies, company lookup, URLs, live signals, or just-founded startups.
startUrls	array	empty	Dealroom company, list, or landscape URLs.
searchQuery	string	empty	Search term for public market-map discovery.
marketmapIds	array	landscape-53885	Market-map ids or Dealroom list URLs.
companyNames	array	OpenAI, Anthropic	Company names to resolve through public Dealroom indexes.
includeNews	boolean	true	Mark whether news metadata is available where lookup results expose it.
includeSentiment	boolean	true	Mark whether sentiment metadata is available where lookup results expose it.
sortBy	string	source	Source order, company name, newest launch year, or highest funding first.
maxItems	integer	20	The only finite default cap. Set 0 for unlimited.
limit	integer	empty	Compatibility alias for competitor-style inputs.
maxPages	integer	empty	Unused reservation field (this actor has no page walk — every mode fetches each target in one request). The run stops at Max items.
pageSize	integer	50	Page size for market-map search and signal modes.
proxyConfiguration	object	Apify datacenter	Dealroom feeds usually work with datacenter or direct routes. Residential can raise cost.
resumeFromRunId	string	empty	Continue ONE specific interrupted run/dataset without re-returning or re-charging already-collected records.
incrementalMode	boolean	false	Remember this search across separate scheduled runs and classify each record NEW / UPDATED / UNCHANGED / REAPPEARED / EXPIRED.
stateKey	string	empty	Name an incremental-mode monitoring campaign, or leave empty to derive one automatically from the mode/search/filter/toggle settings.
emitUnchanged	boolean	false	Incremental mode only. Also return (and bill) records that have not changed since the last run.
emitExpired	boolean	false	Incremental mode only. Also return (and bill) records that were previously tracked but are no longer found — only after a complete scan.

Resume & recurring updates

Two distinct workflows:

• Resume (resumeFromRunId) continues ONE specific interrupted run — paste a previous run or dataset id and this run appends only records not already collected there.
• Incremental mode (incrementalMode) is for scheduling this same search again and again (daily/weekly). The actor remembers state itself in a key-value store keyed on stateKey (or an automatic hash of mode/search/marketmap ids/company names/sort/enrichment toggles — excludes maxItems/limit/maxPages/pageSize and any mcp*/notion* export field, so raising a cap or adding a connector never starts a new baseline). Each record is classified against that state:
  • NEW — never seen before under this state key.
  • UPDATED — a real content change since last seen (changedFields lists what changed).
  • UNCHANGED — identical to last seen; suppressed (not returned or billed) unless emitUnchanged is on.
  • REAPPEARED — was previously EXPIRED and is back.
  • EXPIRED — tracked previously but missing this run. Only emitted when the run proved a complete scan (no source was skipped, no unexpected payload shape, maxItems did not truncate collection, and resumeFromRunId was not used) — otherwise "not found" is indistinguishable from "not reached yet", and previous state is left as-is. If more expired candidates exist than remaining maxItems budget allows, the excess is deferred (not emitted, state left untouched) for the next run.

changeType, changedFields, firstSeenAt, and lastSeenAt are only present on records in incremental-mode runs.

The identity key reuses the actor's existing dedup key (type + id/url/company_name), which is stable across runs for market-map companies, lookups, and just-founded startups (all backed by a stable upstream id/path/slug). For live signals — which carry no upstream id at all — the key is derived from the signal's title (stable content, e.g. "Furiosa AI raised $671M"), not the feed's array position.

Because this actor's maxItems cap binds at collection time (before any classification happens), and suppressed records are never backfilled with additional real ones, a recurring incremental run against a stable, unchanged source genuinely costs less than the first run — fewer records are pushed and billed, nothing refills the suppressed slots.

Change detection ignores only per-run scraping artifacts that are not business data — scrapedAt, the internal scrape_metadata.scraped_time/results_page_index/result_index labels, and list ORDER (a reshuffle with the same members is not a change). Every business field — funding, valuation, employee count, growth stage, news_available/sentiment_available, and news_and_signals.signal_time (a coarse, stable fact about a funding round, not a per-request timestamp) — is compared and can trigger UPDATED.

Example input

{
  "mode": "urls",
  "startUrls": [
    "https://app.dealroom.co/lists/53885",
    "https://app.dealroom.co/companies/openai"
  ],
  "maxItems": 10,
  "sortBy": "name"
}

Output

Each item contains the competitor-compatible envelope fields type, id, and url, plus normalized company groups:

• company_identity
• market_and_product_profile
• growth_and_traction
• funding_and_financials
• investors_and_ownership
• locations_and_operations
• web_and_social_presence
• news_and_signals
• source_record

The actor preserves the original public source object in source_record so downstream users can access source-specific fields without waiting for schema updates.

In incremental-mode runs, each record also carries changeType, changedFields, firstSeenAt, and lastSeenAt — see Resume & recurring updates.

MCP connector export

The optional MCP connector inputs send a concise copy of each saved record into your connected apps. Authorize a connector under Apify Settings and select it in mcpConnectors. For Notion, also provide notionParentPageUrl.

Connector export is a side channel only. The complete JSON record always stays in the Apify dataset, and connector failures do not change scraping results.

Connection guidance

The default route is designed for Dealroom feeds and normally does not need residential proxy traffic. If a direct or free-plan environment is rate-limited, enable Apify Proxy. Residential proxy selection may increase cost and should be used only when repeated source errors indicate it is necessary.

Pricing

This actor is configured for pay per event with a lower per-record headline price than the referenced Dealroom competitor:

• Actor start: $0.08
• Dataset item: $0.002 per saved record
• News and sentiment enrichment: $0.0008 — charged only for a saved record that actually gained a real news_available or sentiment_available signal (not merely for having the toggle on). A record whose lookup returned no news or sentiment is not charged.

Notes

Dealroom public endpoints can expose a capped sample for some market maps. The actor reports the source payload in source_record so users can inspect source totals, returned counts, and cap notes.