Clay

What is Clay?

Clay is a data-rich, automated platform designed for GTM engineering, sales, marketing, and revenue operations teams. It combines corporate and contact search, multi-data source query capabilities, AI-driven insights, signal monitoring, CRM synchronization, and outreach tasks within tables or workflows.

Unlike tools that only offer databases of fixed contacts, Clay functions more like an orchestrable infrastructure for data and growth. Users can connect their own data, as well as providers from the Clay Market and third-party APIs, and then use conditions, formulas, agents, and actions to determine how to process each row of data.

A one-sentence summary

Clay assists teams in locating target companies and contacts; it draws on multiple data sources to gather information such as email addresses, phone numbers, and company details. Then, Claygent is used to analyze this data, assign scores, and generate personalized content, with the results being synchronized to CRM systems, email sequences, or advertising audiences.

Main functions

Company and contact search

Users can create target lists based on criteria such as industry, region, company size, job title, financing status, technology stack, and other factors. Audiences are also used to combine first-party and third-party data, enabling more extensive searches for companies and contacts.

Rich multi-provider waterfall

Waterfall queries multiple data providers in the specified order, and once it finds data that meets the requirements, it stops processing that particular branch. The pricing page mentions over 150 data sources, while the website’s market page cites more than 200 suppliers; these figures may reflect different ranges of calculation.

• Find and verify the work email, phone number, and profile information.
• Add corporate details such as company size, industry, financing, technology stack, and recruitment.
• Sort the queries by supplier coverage, cost, or region.
• Skip subsequent calls when the field already has a value or the confidence level is sufficient.
• Using your own supplier’s API key can reduce the consumption of Data Credits, but it will still be counted as Actions.

Claygent AI research

Claygent is an intelligent web research agent that can access websites, extract information, and generate custom data points based on user instructions. It is suitable for researching company strategies, products, hiring practices, competitors, or any other information that cannot be obtained directly from standard databases.

The research results are entered into the table columns to facilitate subsequent scoring and routing. Changes to the website, access restrictions, and ambiguous evidence can still lead to errors; therefore, important conclusions need to be recorded through specific channels and checked manually.

AI content and data processing

The AI actions can be used for web research, text conversion, classification, content creation, and image generation. The AI formula generator, API configuration generator, and error debugging tools assist non-developers in setting up logic and connections.

• Convert call transcripts, web pages, or free-text into structured fields.
• Classify and score accounts or contacts based on the ideal customer profile.
• Generate personalized outreach snippets by combining company and contact data.
• Use natural language to generate Clay formulas and conditional statements.
• Choose models from OpenAI, Anthropic, Google, and other providers.

Signals purchase signals

Signals allows for tracking changes such as job transitions, promotions, new employees, company news, and social activities. Growth adds Web intent signals and Webhook triggers, making it suitable for turning static lists into continuously updated, trigger-based workflows.

Synchronization between CRM and data warehouse

The Growth solution offers automatic CRM synchronization and enrichment, and it can also be synchronized with data warehouses. The typical process involves reading old records, filling in missing fields, removing duplicates or assigning scores, and then writing the validated results back to the sales system.

Clay Sequencer

Clay Sequencer is a built-in email outreach tool that allows users to create sequences using table fields and AI snippets. Teams can also send data to third-party email platforms, without being forced to use the built-in sending capabilities.

Workflows and Functions

Workflows consist of triggers, branches, and actions to create graphical processes focused on individual records; they represent a new approach different from the traditional tabular layout. The developers currently describe it as an alpha version, and functional changes as well as rollback options should be taken into account for practical use.

Functions allow common search operations, enhancements, and business logic to be encapsulated and reused across various workflows. Launch enables the reuse of such functions; GTM engineers can hand over proven, complex procedures to sales representatives so that they can use them securely.

Advertising Audience and Execution

Clay Ads enables the synchronization of a targeted and filtered audience to advertising platforms such as LinkedIn, Meta, and Google. Growth does not impose any restrictions on the advertising audience; however, the actual matching rate, the minimum size of the audience, and the platform-specific policies are still determined by the respective advertising channels.

Sculptor Dialogue Assistant

Sculptor is a conversational assistant used for creating tables, analyzing current data, and identifying issues. It reads the existing tables and the business context to help generate formulas, understand errors, and optimize procedures.

Complete workflow

1. Define the ideal customer profile, exclusion criteria, and fields that need to be completed.
2. Import the initial list from Clay searches, CSV, CRM, data warehouses, or Webhooks.
3. First, clean the domain name, name, and company fields, and set rules for removing duplicates.
4. Configure emails, phones, and corporate information for cascading queries based on coverage rate and cost.
5. Use conditional logic to enrich only those fields that are missing or have low confidence levels.
6. Public information that makes it difficult to structure Claygent or AI motion studies.
7. Scoring and grouping are performed based on fitness, intent signals, and data integrity.
8. Synchronize qualified records to CRM, sequences, advertising, or other execution tools.
9. Monitor Actions, Data Credits, error rates, and response results, then adjust the process.

Usage tutorial

Create the first lead enrichment table

1. Register a workspace and start with a template or a blank table; import 10 to 20 test records first.
2. Ensure that each row contains at least the company domain name, the contact’s name, or some other stable matching key.
3. Add a step to verify the company details, ensuring that the name, domain name, and industry are correctly matched.
4. Add work email addresses in sequence and enable the verification step.
5. Set it to continue searching only when the email box is empty, in order to avoid repeated charges.
6. Add a Claygent research field, and require brief conclusions along with verifiable evidence.
7. Check the cost and errors of each step, then expand to a complete list.

Build upon trigger-based CRM to enhance its functionality.

1. Connect to CRM in the testing environment, and select a specific object along with a limited number of fields.
2. Set trigger conditions such as creation, update, or field becoming empty.
3. After reading the records, duplicate entries are removed to prevent the creation of a second company or contact.
4. Data source and AI analysis are invoked only for missing or expired fields.
5. Write the results to a temporary field for initial review by the operations staff.
6. Confirm the accuracy and permissions before writing to the official fields.
7. Record each synchronization, exception, and rollback, and set usage alerts.

Control Actions and Data Credits

1. Before running, check the Action and Data Credits costs displayed for each action.
2. Use a filter to narrow down to the records that truly need to be processed.
3. Skip enrichment when the field already has a value or the data is fresh enough.
4. Limit the new process to 10 to 50 lines in order to calculate the cost of a single order completely.
5. Turn off the unnecessary Auto-Update and the continuously running AI column.
6. Teams that have supplier contracts can connect their own API keys and compare the actual costs.
7. Regularly check the Usage Dashboard and table history to identify abnormal usage patterns.

Which users are it suitable for

• GTM Engineer: Building a growth system that integrates various data sources, AI, CRM, and outreach tools.
• Revenue Operations Team: Cleansing CRM data, filling in missing fields, monitoring signals, and maintaining data quality.
• Sales Development Team: Identifies contacts, verifies their details, and creates personalized outreach materials.
• Growth and Marketing Team: Creating ABM lists, advertising audiences, and routing incoming leads.
• Startups: Quickly testing target markets and sales strategies in the absence of a large data team.
• Recruitment team: Researches companies and candidates, but must comply with recruitment and privacy regulations.
• Developer: Connect internal applications via public APIs, CLI, and Agent Plugins.

Typical use cases

• Construction of the total potential market: Search based on company and individual criteria, with gradual expansion.
• Email and phone waterfall: Sequentially contacting multiple suppliers to increase coverage.
• Inbound lead processing: Upon receiving a form, it is automatically completed, rated, and assigned to a sales representative.
• CRM cleaning: Identifies missing, expired, or duplicate records and fixes them in bulk.
• Position change monitoring: Follow-up is initiated upon detecting that the target contact has been promoted or joined a new company.
• Account research: Use Claygent to extract information on products, competitors, hiring, and strategy.
• Personalized outreach: Generate email segments based on verified data and arrange them in a sequence.
• Advertising audience synchronization: Dynamically push highly relevant accounts to advertising platforms.
• Reactivating lost opportunities: Using CRM history and new signals to find reasons for re-engaging.

Product advantages

• Put search, enrichment, AI research, orchestration, and execution in the same data workspace.
• The waterfall query enables teams to combine multiple suppliers, rather than relying on a single database.
• Conditions, formulas, and row-by-row execution allow for precise control over costs and data pathways.
• It supports using custom API keys, allowing teams that already have data contracts to continue using them.
• Claygent can provide web research fields that are not available in standard corporate databases.
• Growth supports complex integrations such as CRM, data warehouses, external APIs, and Webhooks.
• Public APIs, CLI, and Agent Plugins enable technical teams to build within coding agents.
• The free plan and 14-day trial allow you to test the process and data quality first.

Usage restrictions and precautions

• Clay uses both Actions and Data Credits as measurement criteria; it is not sufficient to consider only the monthly subscription fee.
• The cost per piece of data from third parties or for AI research varies, and complex processes can quickly deplete the allocated quota.
• Using your own API key still typically incurs a cost for each action that is enhanced in this way.
• Actions are reset at the end of each cycle and are not carried over, while Data Credits can be carried over under certain conditions.
• One-time Data Credits come with a 30% premium, making them unsuitable for regular replenishment.
• Free allows up to 200 rows per table, and it does not include a large number of phone numbers.
• The 14-day trial does not include access to a variety of phone features, and most of the trial points expire after the trial period ends.
• External data providers may return errors, outdated information, or contact details that belong to someone else.
• When accessing web pages, Claygent may be affected by page updates, login requirements, anti-scraping measures, and the quality of the sources.
• Automatic outreach must comply with anti-spam, privacy, unsubscribe, and target platform rules.
• Workflows are currently marked as alpha, and the Agent Plugin is also in the open testing phase.
• The page showing the public API keys still carries a beta label; those integrating it in production should pay attention to any changes in version and permissions.
• Before automatically writing to CRM, it is necessary to remove duplicates, test the fields, conduct audits, and prepare for rollback.

Prices and packages

The prices listed below were verified on August 21, 2026, and are expressed in US dollars. The official website offers plans named Free, Launch, Growth, and Enterprise; the annual pricing is lower than the monthly pricing. Actions and Data Credits can be upgraded in certain plan levels, and the final bill will be based on the details displayed on the workspace billing page.

Key differences between packages

Rules for trials, transfers, and additional purchases

• The 14-day trial offers 1,000 Data Credits, along with access to Webhook, CRM, Sequencer, and REST API functionalities.
• During the trial period, a large number of phone numbers are not available.
• The unused Data Credits for the monthly plans Launch and Growth can accumulate up to twice the amount allowed per month.
• When renewing an annual plan with the same or higher commitment, up to 15% of the purchase quota from the previous year can be carried over.
• Actions are reset on a periodic basis; they are not carried over, nor can they be purchased separately on a temporary basis like Data Credits.
• One-time Data Credits cost 30% more than regular plans, and are subject to a cap on carryover.
• After the downgrade, the data remains intact, but the functions are limited by the new plan, and the balance of points is adjusted according to the new maximum limit.

How are Actions and Data Credits billed?

Action refers to a task carried out by Clay on platforms such as those for advanced research, data transmission, or synchronization; each completion of such a task is counted as 1 unit. Data Credit is used to purchase third-party data or AI models, with the amount required varying depending on the field, supplier, and model used.

Platform support and integration

Clay is primarily used through a browser; no full native clients for ordinary users on Windows, macOS, iOS, or Android have been found. Technical users can also access its authorized functions from external tools via public APIs, CLI, Agent Plugins, and MCP.

API, CLI, and Agent Plugin

Public API

Clay offers a Public API that enables backend tasks, internal tools, queues, and product features to access Clay’s data and processes. API keys should be stored only on the server; they must not be included in browser code, mobile applications, public repositories, logs, or analysis tools.

Agent Plugin and CLI

The official Agent Plugin bundles the Clay API, CLI, MCP, and various skills together and makes them available to coding agents with a Shell interface such as Claude Code, Codex, and Cursor. It enables searching for companies and individuals, running functions, querying tables, and creating or triggering GTM workflows using natural language.

The Agent Plugin is currently in the open beta phase; according to the official statements, it will be available in all solutions as well as in the older self-service versions by the end of 2026. The cost of using it is the same as that in the regular products. It is not suitable for ordinary chat interfaces that do not support CLI execution.

MCP for Reps

Clay MCP enables sales representatives to access the data and functions maintained by the operations team, using AI-powered tools. Its focus is on providing controlled GTM capabilities to frontline employees, rather than allowing each salesperson to modify the underlying data structures on their own.

Privacy and data security

Clay Labs Inc is the entity that operates this service; it handles account information, the enterprise and contact data that is imported, table contents, integrated data, and usage logs. The customer remains the data controller and must ensure that there is a legitimate basis for importing, enriching, and sharing such data.

According to the official AI documentation, customer data is not used to train the models of Clay or its AI providers; agreements have been made with providers such as OpenAI, Anthropic, and Google to prohibit such training. Only when a user explicitly initiates an AI action are the selected table fields sent to the corresponding model.

• Customer data is primarily processed in the AWS US-East region in the United States.
• Static data and data in transit are encrypted using industry-standard methods, with TLS 1.2 or 1.3 being employed for transmission.
• The authorities list compliance with SOC 2 Type II, ISO 27001, GDPR, and CCPA.
• As the data processor, Clay is responsible for handling the data, while the user decides which columns the AI will see as well as the access rights to the workspace.
• Users can delete records, tables, or workspaces; once a workspace is deleted, its data is removed after 30 days.
• The privacy policy permits the retention of necessary information for legal, anti-fraud, and service operation purposes.
• When using third-party data providers or your own API keys, it is also necessary to review the privacy policies of the respective services.
• Marketing lists should have provisions for data minimization, retention periods, and procedures for requests to unsubscribe or delete data.

GitHub and open source

Clay’s official development documentation points to the public repository for Agent Plugins under the clay-run name; this repository contains instructions regarding skills, configurations, and CLI integration for Claude Code, Codex, and Cursor. The repository page does not indicate a specific open-source license, so just because it is publicly accessible does not mean that free copying, modification, and redistribution are allowed.

The Clay business platform, data market, Claygent, Audiences, Workflows, and Sequencer do not have their complete core source code made public; therefore, these products should be classified as non-open source. Other libraries with similar names related to Clay, as well as basic Earth modeling tools and third-party CLI tools, should not be confused with this product.

Basic information

Recommendation score

The recommendation score is 4.6 out of 5. Clay is suitable for mature GTM teams that need to integrate data from multiple suppliers, as well as for AI research, CRM, and distribution channels; it is particularly appropriate for organizations where staff are constantly working to optimize workflows.

The main drawback is the high learning cost associated with double billing; the pricing is significantly higher than that of simpler listing tools. Moreover, it is still up to the user to ensure data accuracy and compliance with marketing regulations. Small teams should first use trials and a limited amount of data to determine the total cost per entry.

Frequently Asked Questions

Is Clay free?

There is a free plan that includes 500 Actions and 100 Data Credits per month; there is no limit on the number of seats or tables, but each table can have up to 200 rows, and phone support is not provided. It is suitable for learning and small-scale testing.

How much is Clay’s lowest priced subscription plan?

The monthly cost for Launch starts at $185; the annual plan amounts to approximately $167 per month. The basic quota includes 15,000 Actions and 2,500 Data Credits per month, with the annual plan offering the corresponding annual limits.

What is the difference between Actions and Data Credits?

Actions are used to measure the work carried out on the platform by Clay, while Data Credits are used to purchase third-party data or AI capabilities. A single action typically consumes 1 Action unit, along with a varying number of Data Credits.

Is the unused quota carried over?

Actions are not carried over. For monthly plans Launch and Growth, Data Credits can accumulate up to twice the amount permitted for that month; when renewing on an annual basis with a plan of the same level or a higher one, up to 15% of the annual purchase amount can be carried over.

Does Clay offer a trial version?

A 14-day trial period is provided, along with 1,000 Data Credits; in addition, features such as Webhooks, CRM, Sequencer, and REST API are available. A wide range of phone numbers is not accessible during the trial period.

Can I use my own data provider’s API key?

Yes, some enhanced versions support built-in API keys, which can help reduce the cost associated with Clay Data Credits. The platform will still carry out the orchestration processes; therefore, each enhancement typically requires one Action, with the supplier charging extra for this service.

Does Clay offer a public API?

Public APIs are provided, in addition to CLI, Agent Plugins, and MCP. API keys must be stored on the server, and the available functions and access levels depend on the current plan and beta status.

Is Clay open source?

Commercial platforms are not open source. The official Agent Plugin repository is accessible to the public, but no explicit license is indicated there; therefore, it cannot be assumed that the entire product or repository is subject to a standard open source license.

Will Clay use customer data to train AI?

The official AI documentation clearly states that this will not happen, and it mentions that AI suppliers are bound by contracts that prohibit them from using customer data to train models. Users still need to decide which columns can be sent to the AI and to review any third-party integrations.

Is Clay suitable for individual users?

Individuals can use Free for learning purposes, but Clay’s main value lies in handling complex GTM data and workflows. Those who have only a few contacts or rely on simple email inquiries may find it difficult to derive sufficient value from the Launch version.

Is it necessarily compliant to send cold emails from Clay?

Not necessarily. The tool provides data and sending capabilities, but it does not replace the sender’s legitimate basis for sending messages, identification details, mechanisms for unsubscribing, data deletion options, or the obligations related to anti-spam measures in the target regions.

Summary

Clay combines lead search, multi-source data, AI research, signals, CRM, and distribution channels to create an orchestrable GTM system. It is ideal for teams that are willing to turn data processes into scalable products through the use of conditions, testing, and continuous monitoring.

Before making a purchase, it is necessary to verify the coverage rate, error rate, and the cost per record using 10 to 50 real records, before deciding on whether to opt for the Launch, Growth, or Enterprise version. Once the system is in use, it is important to manage issues related to points, data access rights, external compliance requirements, and the risks associated with automatic data writing back.