Company Tech Stack Detection: BuiltWith Alternative

GTM tech stack detection for Clay. Flat boolean output. No JSON parsing required.

---

What's Inside

• Ready-to-Run Examples
• What's New
• Overview
• Who this is for
• Input
• Output
• How signal strength is derived
• Clay Table Template
• Pricing
• Error Handling
• Known Limitations
• Issues and Feature Requests
• Use with MCP (AI Agents)
• Mamba Labs GTM Actor Fleet

Ready-to-Run Examples

Each example is a pre-configured version of this actor for a specific use case. Click any link to open it in the Apify Console and run it immediately.

Example	What it does
Bulk Tech Stack Lookup: BuiltWith Alternative	Enrich a list of domains with detected CRM, sequencer, and marketing automation tools.
Detect Sales Sequencer: Outreach vs Salesloft	Detect which sales sequencer a company runs from its website.
Detect a Company CRM From Its Website	Check whether a company runs HubSpot, Salesforce, or another CRM from its public website.
Find Companies Using HubSpot or Marketo	Segment a list of domains by detected marketing automation platform.
Fresh Tech Stack Re-Scan (No Cache)	Force a fresh detection with no cached result, to catch a recently added or removed tool.

Looking for a different configuration? Open the Input tab to build your own.

What's New

• Pricing reduced. $0.015 per result, down from $0.07. No subscription, no contract.
• Free tier added. Free Apify-plan users get 15 results per calendar month (resets monthly, counts only successful detections); paid plans are unlimited. Upgrade: https://apify.com/pricing.
• Clay table template and function available. Pre-built table plus a reusable Clay Function. Open the template →
• Detection is actually running. Prior versions shipped with a base Docker image that had no Chromium binary. Every run in production was silently failing at chromium.launch() and returning all-false. v0.2.1 switches the base image to apify/actor-node-playwright-chrome:20 which includes Chromium at the path Playwright expects.
• Detection errors now surface in the output. When the browser fails to launch or every fingerprint lookup errors out, detection_error is set to true and detection_error_reason carries the specific failure (for example timeout, error: net::ERR_...). Production failures no longer hide behind a row of false booleans.
• Input field rename. Primary field is now domain to match Actor 1. Legacy company_domain and url are still accepted for backward compatibility. Existing Clay tables keep working.
• Apollo fingerprint updated. New cookie patterns (dwnjrn, zp__initial_landing_page) match what Apollo's current marketing site actually sets.
• Ground truth regression test. tests/ground_truth.js runs the actor against six domains whose marketing sites carry detectable fingerprints. Five of six must pass before any release push.

---

Overview

Takes a company domain or URL, renders the page with headless Chromium, and returns a single flat row with boolean flags for each GTM tool detected. Drop it into a Clay HTTP Request column and filter on gtm_tool_count > 0 to qualify your list instantly.

---

Who this is for

• Clay users building outbound qualification tables who need GTM tool flags without writing formula columns
• GTM agencies enriching lead lists with tech stack signals at scale
• RevOps teams segmenting accounts by CRM, sequencer, and marketing automation usage

---

Input

Field	Type	Required	Notes
domain	String	Yes	Bare company domain. Example: stripe.com. No https://, no www., no trailing slash.
company_domain	String	No	Deprecated. Still accepted for backward compatibility with existing Clay tables.
url	String	No	Deprecated. Still accepted.
crawl_additional_pages	Boolean	No	Default true. Set to false for homepage-only runs.
domains	Array	No	List of bare domains for batch processing. Takes precedence over domain when provided. Each domain produces its own output row.
batchSize	Integer	No	Domains detected concurrently per wave in batch mode. Default 5, maximum 10. One browser is reused across the batch, so memory stays flat; raise the actor memory only if you increase this well beyond the default.
skipCache	Boolean	No	Default false. A clean detection is cached for 7 days and reused on repeat lookups, skipping the browser launch entirely. Set true to force a fresh detection.

Batch usage

Pass a domains array to score many companies in one run. The actor reuses a single browser across the batch (one Chromium process, one fresh context per domain) and processes batchSize domains at a time, so a 50 domain run does not launch 50 browsers:

{
  "domains": ["stripe.com", "notion.so", "figma.com"],
  "batchSize": 5,
  "crawl_additional_pages": true
}

Each domain is emitted as its own flat row, identical in shape to a single domain run. Free-tier runs process up to the remaining monthly allowance and stop.

Caching

A clean detection for a domain is cached for 7 days in a named Key-Value Store in your own account and reused on repeat lookups, skipping the Chromium launch entirely. This makes re-running the same account list cheap. Error and empty detections are never cached. Pass skipCache: true to force a fresh detection.

---

Output

Every field is present in every row. No nested JSON. Ready to use as a Clay enrichment column.

Field	Type	Description
domain	String	Normalized domain from input (new primary key)
company_domain	String	Same value as domain. Present for backward compatibility.
crm_detected	String	hubspot / salesforce / marketo / pardot / none
seq_tool_detected	String	outreach / salesloft / apollo / instantly / intercom / drift / none
uses_hubspot	Boolean	HubSpot CRM or Marketing Hub detected
uses_salesforce	Boolean	Salesforce detected
uses_clay	Boolean	Clay detected
uses_apollo	Boolean	Apollo.io detected
uses_outreach	Boolean	Outreach detected
uses_gong	Boolean	Gong detected
uses_zoominfo	Boolean	ZoomInfo detected
uses_drift	Boolean	Drift detected
uses_intercom	Boolean	Intercom detected
uses_marketo	Boolean	Marketo detected
marketing_automation_detected	String	marketo / pardot / hubspot / none
gtm_tool_count	Integer	Count of true boolean fields
tech_stack_signal	String	high / medium / low
detected_tools	Array	One object per detected tool: name, category, confidence, evidence. Additive structured view of the flat booleans; empty array when nothing is detected.
detection_error	Boolean	True if every fingerprint call returned a browser error or timeout
detection_error_reason	String	Specific error reason when detection_error is true (for example timeout, error: ENOENT). Null otherwise.
run_date	ISO Date	Timestamp of the run

Sample output:

{
  "domain": "hubspot.com",
  "company_domain": "hubspot.com",
  "crm_detected": "hubspot",
  "seq_tool_detected": "none",
  "uses_hubspot": true,
  "uses_salesforce": false,
  "uses_clay": false,
  "uses_apollo": false,
  "uses_outreach": false,
  "uses_gong": false,
  "uses_zoominfo": false,
  "uses_drift": false,
  "uses_intercom": false,
  "uses_marketo": false,
  "marketing_automation_detected": "hubspot",
  "gtm_tool_count": 1,
  "tech_stack_signal": "medium",
  "detected_tools": [
    {
      "name": "HubSpot",
      "category": "CRM",
      "confidence": "high",
      "evidence": "script src match: /js\\.hs-scripts\\.com/"
    }
  ],
  "detection_error": false,
  "detection_error_reason": null,
  "run_date": "2026-04-06T09:03:47.744Z"
}

detected_tools[] is additive: every flat boolean and string field is unchanged, so existing Clay tables keep working. Each entry carries:

• name: display name, for example HubSpot, Salesforce, Marketo.
• category: one of CRM, MAP, Analytics, Chat, CDP, ABM, Ads, A/B Testing, CDN/Framework, Other.
• confidence: high for an exact script or cookie match, medium for a JS global, low for a heuristic HTML match.
• evidence: the signal that triggered the match, for example script src match: /js\.hs-scripts\.com/.

Clay gate column: gtm_tool_count > 0

---

How signal strength is derived

Strength	Criteria
high	CRM detected AND sequencer detected AND 3 or more total tools
medium	CRM or sequencer detected, 1 to 2 total tools
low	At least 1 tool detected, no CRM and no sequencer

---

Clay Table Template

A pre-built Clay table template is available with the enrichment column, formula columns, and views ready to use.

Open the Clay template →

The template includes:

• A Run Apify Actor enrichment column wired to mambalabs/gtm-tech-stack-signal-scraper
• Eight formula columns for CRM, sequencer, and tool detection flags
• A qualifier column for downstream gating

A reusable Clay Function is also available: Open the function →. Use the function when you want updates to the detection logic to propagate automatically. Use the template when you want a static copy to modify.

Manual setup

One flat row per domain, Clay-ready. Add this actor as a Run Apify Actor enrichment column in any Clay table with a company_domain column.

Input mapping:

• Apify Actor: mambalabs/gtm-tech-stack-signal-scraper
• Input Data: {"domain": "[company_domain pill]"}
• Run condition: company_domain is not empty

Output fields map directly to Clay formula columns using dot notation:

• {{actor_output}}?.uses_hubspot
• {{actor_output}}?.uses_marketo
• {{actor_output}}?.uses_intercom
• {{actor_output}}?.crm_detected
• {{actor_output}}?.seq_tool_detected
• {{actor_output}}?.gtm_tool_count
• {{actor_output}}?.tech_stack_signal
• {{actor_output}}?.detection_error

Gate downstream tables on gtm_tool_count > 0 or tech_stack_signal = high to focus on companies running real GTM infrastructure.

---

Pricing

$0.015 per result. Pay per event, billed per domain processed. No subscription, no seat license, no contract.

Tier	Discount	Per result	Per 1K results
Free (no plan)	0%	$0.015	$15.00
Starter (Bronze)	~5%	$0.0143	$14.30
Scale (Silver)	~10%	$0.0135	$13.50
Business (Gold)	~15%	$0.0128	$12.80

Volume	Cost
100 domains	$1.50
1,000 domains	$15.00
10,000 domains	$150.00

We recently reduced pricing across the Mamba Labs actor suite to bring more value to GTM teams running enrichment at scale.

Free tier vs paid usage

Plan	Limit
Free Apify plan	15 results per calendar month (resets on the 1st)
Any paid Apify plan	Unlimited results, no cap

Free-plan usage is counted cumulatively across all of your runs in a month, not per run, so each single-domain run draws from the same monthly allowance. Only successful detections count against the allowance; a run that errors out (bot-block, timeout) does not consume a result. How it behaves when you hit the limit:

• Within the monthly allowance: the domain is detected and returned normally.
• Already at the limit: the run exits immediately with no result and an upgrade message in the run status, before any browser cost is incurred.

Upgrade to any paid Apify plan for unlimited detection: https://apify.com/pricing. Paid-plan runs are never capped and incur no usage tracking.

---

Error Handling

The actor is built to always emit a usable row. It never crashes mid-batch on a single bad domain, and it surfaces real browser failures so you never mistake a failed run for a confirmed negative.

Condition	Behavior	Output
No domain or domains provided	Early exit before any browser launch	Run throws Provide input.domain (single) or input.domains (array), no rows pushed
Free-tier monthly quota already reached	Quota is checked up front, before the Chromium launch, so no browser cost is incurred	Run exits immediately with an upgrade message in the run status, no rows pushed
Page navigation times out (15s) or the browser errors	Detection returns all-false for that domain rather than throwing	Row with detection_error: true and detection_error_reason set to timeout or error: ...
Browser fails to launch for the whole batch	Every domain in the batch is marked failed	Each row has detection_error: true and detection_error_reason carrying the launch error
Additional page (pricing/product) crawl fails	Failure is logged and skipped; homepage signals are still used	Normal row built from the homepage detection, no error flag raised
Bot-protected site returns a valid 200 stub	No error is thrown because the browser got a real response	Row with all-false flags and detection_error: false (see Known Limitations)
Unexpected per-row exception	The row is caught, an empty record is pushed, and the batch continues	Row built from buildEmptyRecord with detection_error_reason of main_catch: ...

---

Known Limitations

Bot-protected domains may return negative results. Sites fronted by Cloudflare Enterprise, Akamai Bot Manager, or PerimeterX often serve a challenge page or stub to headless Chromium requests from a datacenter IP. The detector navigates successfully, captures an empty stub HTML, and returns all-false with detection_error: false (no error was actually thrown, the browser got a valid 200 response). Known examples include salesforce.com, zoominfo.com, and marketo.com. If you need those, flag the rows for manual review. A stealth-mode build with residential proxies is a candidate for a future coverage-expansion release.

Backend-only tools with no web pixel are out of scope. Clay, Salesloft, Instantly, Lemlist, and server-side Segment leave no public client-side signal on a customer's website. The fingerprints file marks these with detectable: false by design. Do not expect uses_clay: true even on companies you know use Clay. This is documented, not a bug.

Vendor marketing sites reflect their marketing stack, not their product. outreach.io runs Marketo on its marketing site. apollo.io runs HubSpot plus Intercom. gong.io runs Marketo. drift.com runs Marketo post-Salesloft-acquisition. The actor reports what the page actually uses, which is the correct outbound-qualification signal. "This company uses Outreach" is not inferable from "outreach.io's marketing site loads Marketo scripts," and we do not hard-code self-domain shortcuts.

Tools loaded only behind authenticated pages will not be detected. The actor crawls the homepage and up to two public pages (pricing, product). Dashboards, customer portals, and auth-gated paths are unreachable.

Heavily obfuscated CDN routing can reduce script src accuracy. Cookie names and JS global variables are checked as primary signals to partially compensate, but companies that fully mask outbound requests through Cloudflare Workers or Fastly with custom domains may still slip through.

LinkedIn Sales Navigator is not detectable and was removed in v0.2.0.

---

Issues and Feature Requests

Found a bug or want to request a feature? Open an issue in the Issues tab above. Response time: typically within 48 hours.

---

Use with MCP (AI Agents)

This actor is callable as a tool over MCP (Model Context Protocol) by AI clients like Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf, and any other MCP-compatible client. Apify hosts the MCP server at mcp.apify.com, so no custom server code is needed.

What this actor does in an MCP context

Give an AI agent a company domain and it returns a flat row of GTM tech stack signals: which CRM, sequencer, and marketing automation tools the company runs on its marketing site. Useful for outbound qualification, account research, and ICP scoring inside an agent loop.

Example prompts:

• "What CRM does stripe.com use?"
• "Does notion.so run HubSpot or Salesforce?"
• "Detect the marketing automation stack on shopify.com and tell me if they use Marketo or Pardot."

Supported MCP clients

Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf, and any MCP-compatible client.

Setup: Hosted server (recommended)

Add this entry to your MCP client config. The OAuth variant prompts you to authorize on first connect, so no token sits in your config file.

OAuth (recommended):

{
  "mcpServers": {
    "mamba-labs-tech-stack": {
      "url": "https://mcp.apify.com?tools=mambalabs/gtm-tech-stack-signal-scraper"
    }
  }
}

Bearer token:

{
  "mcpServers": {
    "mamba-labs-tech-stack": {
      "url": "https://mcp.apify.com?tools=mambalabs/gtm-tech-stack-signal-scraper",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
      }
    }
  }
}

Setup: Full Mamba Labs GTM Suite

Load all 6 Mamba Labs GTM actors as tools in a single MCP server entry.

OAuth:

{
  "mcpServers": {
    "mamba-labs-gtm-suite": {
      "url": "https://mcp.apify.com?tools=mambalabs/gtm-hiring-signal-scraper,mambalabs/gtm-tech-stack-signal-scraper,mambalabs/b2b-buying-signals-hiring-tech-stack-intent-for-clay,mambalabs/job-board-keyword-signal-scanner,mambalabs/domain-to-linkedin-url-resolver,mambalabs/icp-account-lead-scoring-fit-scorer-0-100-for-clay"
    }
  }
}

Bearer token:

{
  "mcpServers": {
    "mamba-labs-gtm-suite": {
      "url": "https://mcp.apify.com?tools=mambalabs/gtm-hiring-signal-scraper,mambalabs/gtm-tech-stack-signal-scraper,mambalabs/b2b-buying-signals-hiring-tech-stack-intent-for-clay,mambalabs/job-board-keyword-signal-scanner,mambalabs/domain-to-linkedin-url-resolver,mambalabs/icp-account-lead-scoring-fit-scorer-0-100-for-clay",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
      }
    }
  }
}

Setup: Local server (for development and testing)

npx @apify/actors-mcp-server --tools mambalabs/gtm-tech-stack-signal-scraper

Full suite:

npx @apify/actors-mcp-server --tools mambalabs/gtm-hiring-signal-scraper,mambalabs/gtm-tech-stack-signal-scraper,mambalabs/b2b-buying-signals-hiring-tech-stack-intent-for-clay,mambalabs/job-board-keyword-signal-scanner,mambalabs/domain-to-linkedin-url-resolver,mambalabs/icp-account-lead-scoring-fit-scorer-0-100-for-clay

Setup: Dedicated MCP wrapper

Install the dedicated wrapper for this actor and point your MCP client at it:

npm install -g @mambalabsdev/mcp-gtm-tech-stack-signal-scraper

{ "mcpServers": { "mamba-gtm-tech-stack": { "command": "npx", "args": ["-y","@mambalabsdev/mcp-gtm-tech-stack-signal-scraper"], "env": { "APIFY_TOKEN": "YOUR_APIFY_TOKEN" } } } }

Prefer one install for the whole fleet? The Mamba Labs GTM Suite (https://www.npmjs.com/package/@mambalabsdev/mcp-gtm-suite) exposes eleven of these actors as tools in a single MCP server.

What the output looks like

One flat JSON row per domain, well under 2KB. Every field is always present. Key fields:

• domain: normalized input domain
• crm_detected: hubspot, salesforce, marketo, pardot, or none
• seq_tool_detected: outreach, salesloft, apollo, instantly, intercom, drift, or none
• marketing_automation_detected: marketo, pardot, hubspot, or none
• gtm_tool_count: integer count of detected tools
• tech_stack_signal: high, medium, or low
• detection_error and detection_error_reason: surface real browser failures so agents do not treat empty results as confirmed negatives
• Per-tool booleans: uses_hubspot, uses_salesforce, uses_apollo, uses_outreach, uses_gong, uses_zoominfo, uses_drift, uses_intercom, uses_marketo, uses_clay

Billing

MCP runs use the same pay-per-event pricing as direct API or Console runs. Each successful result costs $0.015.

Notes

This actor uses headless Chromium on Apify infrastructure. The MCP client does not need to run a browser locally. Typical run time 5 to 15 seconds.

---

Mamba Labs GTM Actor Fleet

Actor	What it does	Price/result
GTM Hiring Signal Scraper	Detects GTM hiring from career pages (Greenhouse, Lever, Ashby)	$0.05
GTM Tech Stack Signal Enrichment	Detects CRM, sequencer, and marketing automation from a public site	$0.015
GTM Signals Aggregator	Combines hiring and tech signals into one composite GTM score	$0.09
Job Board Keyword Signal Scanner	Scans 5 ATS platforms for roles in any category	$0.05
Domain to LinkedIn URL Resolver	Resolves a domain or name to its LinkedIn URL with firmographics	$0.006
ICP Fit Scorer	Scores a company against your ideal customer profile	$0.05
Domain Deliverability Checker	Audits email deliverability: SPF, DKIM, DMARC, MX, health score	$0.005
Company Firmographic Enricher	Enriches a domain into employee band, industry, HQ, revenue	$0.004
Company Social Presence Mapper	Maps a domain to social URLs and follower counts	$0.015
Company Identity Resolver	Resolves name, domain, or LinkedIn into one canonical identity	$0.007
Company Change-Event Feed	Monitors a domain and returns only what changed since last run	$0.06
Funding & Press Signal Scanner	Scans news and press for funding, exec moves, launches, M&A	$0.03

One-install option: the Mamba Labs GTM Suite MCP server exposes eleven of these actors as tools in a single package. Each actor also has its own MCP wrapper.

All actors: apify.com/mambalabs | Website: mambabuilt.com

Built by Mamba Labs.