Extruct AI
A company research API and web app querying a curated 10-million-company index and the live web through semantic, lookalike and deep search modes, returning every answer with its source for account lists, CRM enrichment and due diligence.
What is Extruct AI?
Extruct AI is a company search and research engine, available both as a web application and as a REST API. It works on two kinds of material at once: a curated index of 10 million companies, and the live web, queried on demand. The stated principle is right answer or no answer: every result comes back with the source it was taken from, which makes each field checkable rather than merely plausible.
Discovery happens through three modes. Semantic Search takes a query in ordinary English and searches the index for intent rather than keywords, favoring recall. Lookalike Search starts from a reference domain or company UUID and returns similar companies. Deep Search runs asynchronously, evaluating candidates against several criteria at once; it returns fewer results, but better qualified. Alongside these, Deep Research produces sourced reports on companies and on people, and AI Tables adds enrichment and scoring columns on Extruct results or on a list the user imports, with a library of reusable columns. Every company is scored from 0 to 100% against the criteria, and rows with a poor fit do not consume credits.
The example shown on the homepage gives the measure of the thing: commercial HVAC contractors in Florida serving hospitals and school districts, 50+ employees, a $5M project in the last twelve months, BIM/Revit capability and actively hiring, five criteria, 2,438 candidates sourced, 112 high-fit matches in 5 minutes 11 seconds.
The publisher released its own benchmark in March 2025: 87.8% overall accuracy against 80.8% for Clay and 77.3% for Perplexity, with Exa excluded for returning over 70% Not Found. The method tested founding year, chief executive and headquarters address plus one sector-specific field, across three segments: plywood, apparel and startups.
A remote MCP server exposes the same search to Claude, Codex and agent tooling through an official extruct-api skill. Two side products complete the picture: Data Room, continuously indexed public datasets, and published market research. Studio is a separate services offering where the team builds data, workflows and internal tools inside the client's organization. Extruct AI, Inc. is a Delaware company run by two co-founders, Dima Persiyanov on engineering and Danny Chepenko on go-to-market, with five published customer case studies: Finom, True Capital, Overdrive, Kewazo and Oftex.
What it does
- Describe a target company profile in plain English and get back a scored list of candidates
- Find companies similar to a reference company with Lookalike Search
- Evaluate candidates against several criteria at once with Deep Search and keep only the best matches
- Enrich an existing list with custom research columns in AI Tables
- Produce sourced research reports on a company or a person with Deep Research
- Retrieve email addresses and phone numbers for target contacts
- Push results to Salesforce, HubSpot, Affinity, Attio or Pipedrive, or plug the search into AI agents through the remote MCP server
When to use Extruct AI / When not to
A quick filter to help you decide if Extruct AI is the right fit.
When to use Extruct AI
- B2B go-to-market and sales teams building account lists on fine-grained criteria that classic filters cannot express
- Investment firms and funds working on deal flow, scouting and due diligence, as in the published Overdrive and True Capital case studies
- Compliance and KYB teams: Finom resolves the official website of every applicant in real time, on 100% of onboarding traffic in production across five countries
- Procurement and sourcing teams verifying suppliers and their certifications, including in low-visibility sectors such as plywood or industrial manufacturing where standard filters fail
- Developers and data teams building on an API-first product capable of hundreds of thousands of agent runs
When not to use Extruct AI
- Teams running HIPAA- or FISMA-regulated workflows: the terms state the services are not tailored to those regimes, and any use that would breach the Gramm-Leach-Bliley Act is barred
- Data brokers and resellers: reselling or redistributing the data in bulk is contractually prohibited, as is using it to build or train a competing database or model
- Organizations bound to European data residency: servers are in the United States and no European hosting region is offered
- Mobile-first users: there is no iOS or Android application, access is through the web app, the API and a Chrome extension
- Teams working outside English: the interface and the queries are English-only, and the service is restricted to users aged 18 and over
How to use Extruct AI
A typical end-to-end flow, from setup to results.
- Create a free account on the web app; nothing has to be installed
- Describe your target profile in plain English, for example "RICS registered property managers in London", "Biotech companies with oncology pipelines and FDA approval" or "Series B companies using Kubernetes in production"; the more detailed the description, the better the match quality
- Let the system break the query down into workable dimensions: location, qualification and business context
- Read the fit score given to each company: 100% is a perfect match, 0% none, 1 to 99% partial
- Use Find More to generate further batches of candidates; two to four iterations usually produce a usable dataset
- Add enrichment columns with an explicit prompt saying what to look for, where to look and how to interpret it, for instance RICS verification status, property types managed, ESG score, recent funding or tech stack, or pick from the ready-made column library
- Find contacts by describing the role you want, such as "VPs of Sales at Series B SaaS companies" or "CTOs at AI startups"
- Push the qualified list to your CRM (Salesforce, HubSpot, Affinity, Attio, Pipedrive) or to an automation flow (n8n, Zapier, Make)
- On the API side, collect the API key issued at sign-up, read the reference on the documentation site and call the search, asynchronous discovery, deep research and table endpoints
- For agent use, connect the remote MCP server once with your Extruct account; the Chrome extension is included on every plan
Pros & Cons
Pros
- Every answer is returned with its source, so results can be verified line by line instead of taken on trust
- A published accuracy benchmark with its method, its segments and its figures, uncommon in this category, including 90.0% on the plywood segment against 67.5% for Perplexity, where classic filters usually fail
- Cost follows relevance: low-fit rows consume no credits, a single credit pool covers every function, and the per-action rate is published
- A permanent free plan of 25 credits per month that already includes API access and the Chrome extension
- API-first design with complete public documentation, a published OpenAPI specification and a public status page
- Numerous CRM and automation integrations, each documented individually
- A serious GDPR setup for a US vendor: Standard Contractual Clauses, a DPA incorporated into the terms, a named Article 27 representative, and retention capped at twelve months after account termination
Cons
- The /dpa/ page referenced and made binding by the terms renders no text when loaded, so the agreement cannot be read as it stands
- Nothing is said about model training on customer data, and no opt-out is documented
- No formal subprocessor list: providers are named in passing inside the privacy policy, with no dedicated or versioned page
- The differentiating features, Deep Search, AI Tables and contact lookup, only start on the Pro plan at $149 per month, and phone lookup at 20 credits per number burns the entire Pro quota in 100 numbers
- Credits do not roll over: the allowance resets every month on all three self-serve plans
- No security certification is displayed, neither SOC 2 nor ISO 27001, and servers are in the United States only with no European data residency option
- A thin company surface: no about page, no contact page, no mobile app, no dedicated legal contact, Team and Enterprise pricing only through a sales conversation, and an accuracy benchmark produced by the vendor on its own test segments
Pricing & Plans
Extruct AI operates on a freemium basis. A permanent free plan grants 25 credits per month at no cost. The lowest paid entry point is the Starter plan at USD 49 per month for 500 monthly credits, followed by Pro at USD 149 per month for 2,000 monthly credits. Team and Enterprise are priced per seat, on quotation only, following a conversation with the founders. All functions draw on a single credit pool with a published per-action rate: 1 credit for a semantic search, 1 for a lookalike search, 2 per Deep Search match, 1 per AI Table cell, 4 per email lookup and 20 per phone number lookup, while rows with a poor fit consume no credits. Credits do not roll over and the allowance resets each month on all three self-serve plans. Upgrades take effect immediately with pro-rated credits, downgrades at the following billing cycle. Payment is by major credit card in self-serve, and NET 30 invoicing is available for team and enterprise customers.
- 25 semantic searches
- 25 lookalike searches
- 25 results per search
- access to the public Data Rooms
- API access and Chrome extension
- 500 semantic searches
- 500 lookalike searches
- 2
- 000 results per search
- public Data Room export
- table and search sharing
- API access and Chrome extension
- 2
- 000 semantic searches
- 2
- 000 lookalike searches
- 2
- 000 results per search
- Deep Search with 1
- 000 matches and 100 results per deep search
- shared workspace
- per-seat pricing
- onboarding on the customer's own workflows
- CRM integration support
- dedicated Slack and email support
Data, GDPR & hosting
A consolidated view of how Extruct AI handles your data.
GDPR overview
GDPR implementation is documented and specific. The privacy policy names the legal bases relied on under the GDPR and UK GDPR (consent, contract, legal obligation and legitimate interests) and lists data subject rights in full: access and a copy, rectification, erasure, restriction, portability, objection, and the right not to be subject to a solely automated decision. Where such a decision has legal or similarly significant effects, Extruct undertakes to inform the person, explain the main factors and provide a simple route to human review. Transfers from the EEA and the UK rely on the European Commission's Standard Contractual Clauses, servers being located in the United States. A Data Processing Agreement at extruct.ai/dpa is incorporated by section 26 of the terms of service dated May 8, 2025 and prevails in case of conflict. Dmitry Persiyanov is the appointed Article 27 representative in the EEA.
Who owns the data?
Extruct AI, Inc. acts as the data controller for personal information belonging to residents of the European Economic Area, and has appointed an Article 27 representative based in Barcelona. Users keep full ownership of what they submit: the terms state that Extruct asserts no ownership over user contributions, nor over the intellectual property or other proprietary rights attached to them. Payment data is handled and stored by Stripe, not by Extruct. Where Extruct processes personal data on a customer's behalf, the Data Processing Agreement published at extruct.ai/dpa governs that processing; it is incorporated into the terms of service dated May 8, 2025 and prevails over them in the event of conflict.
Reuse rights
Extruct collects the identity data users provide, namely names, email addresses, contact or authentication details and phone numbers, and states that it does not process sensitive information. Technical data is captured automatically: IP address, browser and device characteristics, operating system, language, referring URLs and precise or approximate location, the latter being refusable at device level. Signing in through a social account such as Facebook or X transmits the associated profile. AI features are delivered through named third-party providers, Groq, OpenAI and Anthropic, to which input, output and personal data are passed in order to run the service; the declared AI functions cover research, search, natural language processing, text analysis, machine learning models, bots, applications and automation. Google Analytics is used for audience measurement, with an opt-out link. Legal bases under the GDPR and UK GDPR are set out explicitly: consent, contract, legal obligation and legitimate interests. Nothing in the published documents states whether customer data is used to train models, and no opt-out is documented. On the output side, customers may use the company data they retrieve inside the service, but bulk resale or redistribution is prohibited, as is any use aimed at building or training a competing database or model, and access must stay within the published interfaces and the applicable rate and credit limits.
Data retention & training
Hosting summary
Extruct states in section 8 of its privacy policy that its servers are located in the United States, and warns that using the service from elsewhere means personal data may be transferred to, stored in and processed in the United States and in other countries. Network-level observation is consistent with that statement: the site resolves to an anycast node on Google LLC infrastructure (AS396982), geolocated in Kansas City, United States. For transfers out of the European Economic Area and the United Kingdom, Extruct relies on the European Commission's Standard Contractual Clauses, applied both between group entities and between Extruct and its third-party providers. No European hosting region is offered and no data residency choice is exposed to customers. The company displays no hosting or security certification: neither SOC 2 nor ISO 27001 is mentioned anywhere on the site. Buyers under a strict EU-only processing requirement should treat this as a blocking point and raise it with the vendor before contracting.
Things to keep in mind
Risks and trade-offs to weigh before adopting Extruct AI.
- The /dpa/ page referenced by both the terms and the privacy policy renders no text when loaded, so the agreement that governs processing cannot be read before signing
- No documented way to exclude customer data from model training: the subject is not addressed anywhere in the published documents
- Input, output and personal data are passed to third-party AI providers, Groq, OpenAI and Anthropic, in order to deliver the service
- Servers and processing are in the United States only, with no European data residency option
- The tool builds and enriches data on companies and on people: the compliance of downstream use, prospecting and GDPR included, remains the customer's responsibility
- Anti-scraping and fair-use clauses allow suspension without compensation if usage degrades the service or steps outside the published interfaces, credits are lost at month end, and phone lookup costs 20 credits per number
- The accuracy benchmark is produced and published by the vendor on segments it chose, so treat it as directional and validate on your own data; a cited source can also be stale or misread, and automation should not replace judgment on a name you are about to act on
Setup & Integrations
Technical difficulty
Low for interface use, moderate for the API. Creating a free account takes a minute, nothing is installed, and the first natural-language query runs immediately; the Chrome extension is included from the free plan. On the API side, a key is issued at sign-up, authentication is documented and endpoints are REST with a published OpenAPI specification. The only real work is the asynchronous tasks, Deep Search and Deep Research, which require creating a job then polling for the result. Agent use needs one connection to the remote MCP server. CRM onboarding assistance comes with Team and Enterprise plans.
Deployment
Integrations
Supported languages
Behind Extruct AI
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Extruct AI.
Frequently asked questions
Is there a free plan?
What is a credit?
Does searching always consume credits?
What is the difference between Semantic Search, Lookalike Search and Deep Search?
Which plan should I start with?
Do unused credits roll over?
Can I change plan at any time?
What happens when my credits run out?
Is there an API?
Where is the data hosted?
Should you pick Extruct AI?
Extruct AI takes a narrow position and holds it: company research treated as infrastructure rather than as a feature, exposed through an API as readily as through a web app. What sets it apart is not the size of its index but the discipline attached to it: every answer arrives with the source it came from, and the vendor publishes an accuracy benchmark complete with its method, its three test segments and its comparative figures, which is uncommon in a category where claims usually go unmeasured.
The commercial model is equally legible. One credit pool covers every function, a published rate card states what each action costs, low-fit rows are not billed, and the permanent free plan includes API access and the Chrome extension rather than holding them back.
Two reservations temper this. The legal documents lean on a Data Processing Agreement whose page renders no text when loaded, so a customer cannot read the instrument the terms of service incorporate and give precedence to. And nothing addresses whether customer data feeds model training, with no opt-out documented, while input, output and personal data pass to Groq, OpenAI and Anthropic to deliver the service. A further reservation is commercial: the features that justify choosing Extruct over a generic search tool, Deep Search, AI Tables and contact lookup, begin at $149 per month.
The publisher is young: the domain was registered in March 2024 and the first archived capture dates from June 2024. Five named production customers already back the product, across KYB onboarding, investment scouting and industrial sourcing. For teams that need company data they can verify line by line, Extruct AI deserves a run on the free tier, with a direct question to the founders about training and the DPA before any commitment.
- Choosing a selection results in a full page refresh.
- Opens in a new window.