Harmonic uses the AI ​​​​agent Scout to realize startup discovery, market research, network mapping and due diligence automation, covering 35 million companies and 195 million contact records, serving VC and corporate development teams.

Harmonic

Harmonic’s core parameters and statistics

Harmonic is positioned as a "startup discovery engine" and provides a one-stop data platform from company discovery, in-depth research to action exposure for venture capital institutions and corporate development teams. It is not a general AI chat tool, but a professional data infrastructure built around VC investment workflow.

Projects	Public Information
Official positioning	The startup discovery engine
Data coverage	35 million company records and 195 million contact records
Core AI capabilities	Scout AI agent, network map mapping, automated research, intelligent alarms
Delivery form	Web console / REST API / Bulk Data (S3/BigQuery/Snowflake) / MCP Server
Integration method	Chrome extension Zapier, n8n, API, MCP protocol
Security Compliance	SOC 2 Type II, ISO 27001
Home	US (New York Headquarters)
Latest version	Scout AI Agent + MCP integrated version (~2026-06)
Support Platform	Web, API
Financing status	Cumulative US$30 million

Data scale: Harmonic covers the full life cycle data of 35 million companies from registration to IPO, as well as the professional background and relationship network of 195 million people. It is currently one of the most extensive early-stage startup company databases. Data update frequency ranges from real-time (Web/API) to weekly refresh (Bulk Data).

Differentiated positioning: Different from traditional competing products such as PitchBook and Crunchbase, which focus on late-stage companies, Harmonic covers the entire spectrum of companies from the founder's initial registration to the recruitment of CFOs in the IPO stage, and uses the AI ​​​​agent Scout to change the "search-research-reach" link from manual to semi-automatic, reducing the time loss of VC analysts switching between multiple tools.

Harmonic’s users and market recognition

Customer composition: Harmonic has served hundreds of venture capital institutions and corporate development teams. Clients displayed on the official website include Brex, HubSpot, Khosla Ventures, Goldman Sachs, Accel, Anthropic, NEA, Notion, Bessemer, First Round Capital, Menlo Ventures, Salesforce and other well-known investment institutions and enterprises.

Case Verification: The public cases on the official website cover investment institutions at different stages such as Valo Ventures, Bedrock, MVP Ventures, Ridge Ventures, Rillet, Kindred Partners, GTMfund, PostHog, Sorenson Capital, etc. Among them, Tier 1 VC directly quoted "We found 4 out of our last 9 deals on Harmonic", indicating its actual contribution rate as the source of the project.

Investor background: Harmonic’s investors include Craft Ventures, Floodgate, Sozo Ventures, Teamworthy Ventures, Fin Capital, AVG, etc. These funds are also its early customers - using the product first and then investing, indicating that the product itself has been verified by professional users.

Industry benchmarking: Judging from public content, Harmonic is directly benchmarked against PitchBook and Crunchbase, positioning it as an "earlier, more real-time, and more AI-driven" alternative. Its blog systematically compares PitchBook competitors and Crunchbase competitors, demonstrating clear competitive substitution intentions.

Harmonic’s cost advantage

The cost structure is split into three tiers, and there is currently no public price, which is subject to business negotiations.

• C-side/individual users: Harmonic does not provide self-service subscriptions for individual investors or independent angels. The official website does not have a paid package for registration and use, and only provides the "Book a demo" entrance. The threshold for personal use is high and access rights need to be obtained through the enterprise-level sales process.

• Developer/API Access: Provides API Access solutions to support data enrichment, company/person search, and automated integration. The billing model is not public, please refer to the official real-time pricing page. Additional options include financing and investor data (Add-On), codeless CRM synchronization (Add-On).

• Enterprise/Private Deployment: Provides three product lines: Enterprise Console (full-featured web product), API Access (data interface), and Bulk Data (direct delivery from data warehouse, supports S3/BigQuery/Snowflake). Bulk Data contains 35 million company records + 195 million contact records, refreshed every week. Please contact sales team for price.

Hidden Costs: Because the product is targeted at institutional customers, the procurement process involves security review (SOC 2 / ISO 27001 certified), contract negotiations, and annual commitments, making it nearly impossible for an individual or small team to acquire at a reasonable cost. In addition, the data warehouse transmission costs and data governance costs of the Bulk Data solution are borne by the purchaser.

Main functions of Harmonic

• AI agent Scout: Use natural language to describe investment strategies (such as "Looking for vertical players with differentiated advantages in the field of AI infra"). Scout automatically completes market mapping, competitive landscape analysis, team background assessment and growth trajectory judgment. Unlike fixed filters, Scout can understand intent and generate structured research reports, including founding team background analysis, competitive positioning map, financing history and growth assessment.

• Multi-dimensional search and filtering: Supports filtering by dozens of dimensional combinations such as company stage, industry, geographical location, financing status, growth rate, team background, etc. The data granularity covers the entire spectrum from the founder's first submission of company registration to the large-scale recruitment of CFOs in the IPO stage, and the search results can be saved as a monitoring list to track changes in real time.

• Network Relationship Mapping: Map all the team's LinkedIn connections, calendar events, email address books, and CRM records to the database, automatically discovering the warm path of "who knows who". Significantly reduce network blind spots by automatically sending alerts when a former colleague of someone in your portfolio launches a new company.

• Automated monitoring and alarm: Set monitoring conditions for keywords, industries, personnel changes, financing events, etc., and actively push when new matching records are added to the database. For example, monitor signals such as "former FAANG AI researcher leaves to start a business" or "first round of financing occurs in a certain track" to ensure that early opportunities are not missed.

• Research Automation and Export: Generate due diligence reports including company profile, team background, competitive landscape, and financing history with one click. Supports CSV export and Bulk Data warehouse-level data delivery. The report content combines Harmonic’s proprietary data with public information from the entire network, reducing analysts’ time to manually integrate from multiple data sources.

Functional linkage effect: Companies searched by Scout can directly enter the monitoring list → trigger popular path push → generate due diligence report → push to CRM through Chrome extension or API, forming a "discovery → research → reach" relationship instead of an isolated function point. This workflow integration is a key differentiator from traditional databases like Crunchbase.

Harmonic’s model and version evolution

Harmonic's version evolution takes product capability milestones as the main line, gradually iterating from a basic database to an AI-driven end-to-end investment research platform.

Initial database version (~2023)

The first public version, the core capabilities are startup company database and basic search. The data covers millions of early-stage companies, and the seed users are mainly a few VC institutions. The focus of this stage is to verify whether the data quality and update frequency can meet the daily use needs of professional investors.

Search and Monitoring Edition (~2024)

Introducing advanced search filters and custom monitoring alert functions. A new team collaboration workbench is added to support sharing lists and notes within investment institutions. Enterprise customers began to access it on a large scale, and SOC 2 Type II certification was completed in terms of security compliance. At this stage, the data scale has expanded to tens of millions of company records, and basic mapping of network relationships has begun to be supported.

API + Network Mapping Version (~2025)

Open REST API to support data enrichment and search automation. Launched a Chrome extension that allows users to instantly get information about companies or people while browsing the web. The network mapping function has been upgraded to support the import of LinkedIn connections, calendar and mailbox data, and automatically generate an organization-level relationship map. Added Bulk Data service to deliver full data through S3/BigQuery/Snowflake.

Scout AI Agent + MCP integrated version (~2026-06, currently the latest)

Launched AI agent Scout to support natural language-driven market research and investment analysis. Integrated MCP protocol (Model Context Protocol), allowing large language models to directly query the Harmonic database. Supports both Zapier and n8n no-code workflow integration. Publishes the Harmonic Hot 25 Startups quarterly list, which calculates popularity based on aggregated interest signals from thousands of VC institutions. ISO 27001 certification completed.

Harmonic’s technical advantages

Intent understanding mechanism of AI agent Scout: Scout is not a traditional keyword search, but uses a large language model to convert the investment strategy described by the user's natural language (such as "vertical players with differentiated advantages in the field of AI infra") into multi-layer query conditions - industry classification → growth indicators → team background → competitive positioning → financing history, and then comprehensively sort and return the results. This link of "Strategy→Query→Analysis→Report" compresses the manual multi-source investigation that used to take analysts 2-3 hours to minutes.

Real-time data pipeline: Harmonic's data collection covers company registration documents, financing announcements, recruitment information, news, social media and other channels, and is updated in real time through automated pipelines. Compared with traditional databases such as PitchBook that rely on manual input by analysts, Harmonic has a significant advantage in the timeliness of coverage of early-stage companies - it can be included in the database when founders first submit registration, rather than waiting until the financing round is public.

Automatic mapping algorithm for network graphs: By importing the email, calendar and LinkedIn data of institutional members, Harmonic automatically builds an institution-level relationship graph. When a former colleague or alumnus of an investor establishes a new company, the system automatically marks the warm path and pushes it. This ability to "network institutional knowledge" solves the long-standing knowledge management pain point in the VC industry of "information is in the minds of partners and is taken away when they leave."

Native integration of MCP protocol: Harmonic is one of the first commercial data platforms to natively support the MCP protocol. Through MCP Server, large language models can directly call tools such as search_companies, get_company_profile, find_connections, etc. to embed Harmonic data into any MCP-supported AI workflow. This means that users can query startup data directly in clients such as Claude and Cursor without switching to the Harmonic console.

Security Compliance Architecture: SOC 2 Type II and ISO 27001 dual certification, supporting enterprise-level permission isolation. Both API and Bulk Data solutions support data access auditing to meet the compliance requirements of financial institutions. The data will not be used for secondary training of the model, and the data isolation of enterprise customers is guaranteed by contract.

How to use Harmonic

Harmonic is aimed at institutional users, and the usage path is divided into three levels:

Web Console: Obtain an account after booking a demo through the official website. Log in to console.harmonic.ai to enter the search workbench. You can use the Scout AI agent to enter natural language queries, create monitoring lists, view network relationship maps, and generate research results reports. Suitable for daily search and research scenarios of investment teams.

API and Automation Integration: After registering for API access, embed Harmonic data into your own systems via the REST API. API supports company/person enrichment, search and monitoring. For non-technical teams, no-code workflows can be implemented through Zapier or n8n pre-built connectors, such as "When Harmonic monitors a new financing event → automatically write to Airtable and notify Slack".

MCP Server integration: In an AI client that supports MCP (such as Claude Desktop), after adding Harmonic MCP Server to the configuration, the AI ​​assistant can directly call Harmonic data. The typical startup configuration is as follows (subject to the official warehouse README):

{
  "mcpServers": {
    "harmonic": {
      "command": "npx",
      "args": ["@harmonic/mcp-server"],
      "env": {
        "HARMONIC_API_KEY": "<YOUR_API_KEY>"
      }
    }
  }
}

After the configuration is completed, the AI assistant can perform operations such as 'searching for A-round startup companies in the AI infra field', 'querying the background of the founding team of this company', etc. without leaving the conversation interface.

Bulk Data: For engineering teams who need full data, Harmonic regularly delivers data warehouses via S3, BigQuery or Snowflake, supporting large-scale offline analysis and custom model training.

Harmonic product pricing

Harmonic does not provide public pricing, all plans must be quoted through the sales team. Products are divided into three product lines based on delivery form:

Product Line	Applicable Objects	Core Delivery	Billing Model
Enterprise Console	VC investment team	Web full-featured + Scout AI + network mapping + monitoring alarm	Annual subscription, priced by seat
API Access	Developer/Enterprise	REST API + Data Enrichment + Search	Per usage or annual subscription
Bulk Data	Data/Engineering Team	S3/BigQuery/Snowflake Full data refreshed weekly	Annual contract, priced according to data range

Additional Features: Financing and investor data (Add-On), codeless CRM synchronization (Add-On) for an additional fee.

Free Quota: No public free package. All users need to obtain access rights through business processes. Please refer to the official real-time pricing page.

Application scenarios of Harmonic

• Early project source mining: The VC investment team sets monitoring conditions such as "former FAANG researcher quits to start a business", "YC latest batch of companies", "first round of financing in AI track", etc., and Harmonic pushes them in real time when new matching records are added to the database. Compared with the traditional sourcing model that relies on recommendations from personal networks and contacts, the proportion of active discovery has increased from "less than 30%" as deduced to "covering more than 90% of newly registered related companies", significantly reducing the risk of omissions.

• Due Digging Research and Competitive Analysis: Analysts input "assess this company's competitive positioning" through Scout, and AI automatically generates a due diligence report that includes market structure, competitive product comparison, team background, and financing history. Traditionally, a medium-complexity company due diligence report requires analysts to switch between Crunchbase, LinkedIn, news, and recruitment websites for about 4-6 hours; Harmonic compresses the information collection and preliminary screening stages to less than 30 minutes.

• Portfolio Network Activation: After the investment institution imports the relationship network of team members, Harmonic automatically identifies the former colleagues, alumni, joint ventures and other relationships of the founders of the invested companies, and pushes a warm path when these people create new companies. This kind of "relationship-driven sourcing" transforms passive waiting for the Pitch Deck into active discovery, and is especially suitable for mid- to late-stage funds that emphasize "relationship investment."

• Internal Enterprise Innovation Scanning: Corporate strategy/innovation departments use Harmonic to monitor new companies in specific technology tracks and identify potential acquisition targets or technology partners. For example, car companies scan for new battery technology innovations, and financial institutions track emerging fintech players. The Bulk Data solution can directly import data into the internal data analysis platform for customized modeling.

Harmonic is suitable for people

• VC investment team (partner/investment director/analyst): core group. Partners use Scout to quickly scan the track pattern, analysts use search and monitoring for daily sourcing, and the post-investment team uses network mapping to discover follow-up investment opportunities. Unsuitable scenarios: For micro funds that do not rely on systematic sourcing and rely only on personal networks for deal flow, the purchasing ROI may not be significant.

• Corporate Development and Strategy Team (Corporate Development): Used to scan acquisition targets, monitor competitive dynamics, and identify technology cooperation opportunities. Bulk Data solutions are particularly suitable for large enterprises that need to combine startup data with internal data platforms. Prerequisite: Enterprises need to have basic data engineering capabilities to connect to APIs or data warehouses.

• Investment Banks and Research Agencies: Used to prepare market maps, competitive landscape reports and company valuation references. Harmonic covers a wider range of early-stage company data than PitchBook and is suitable for "market analysis before the pre-IPO stage." Unsuitable scenarios: Analysts who need in-depth financial data (such as detailed revenue splits, income statements), Harmonic does not provide financial modeling level data.

• Not applicable to the group: Independent angel investors (no personal package), general researchers in the non-investment field (the data dimension is highly focused on startups), users who need Chinese interface or Chinese company data (Harmonic does not explicitly cover the Chinese market, and the product language is English).

Summary and Outlook

Harmonic's core competitiveness lies in integrating "startup company database + AI research agent + network relationship map" into one package, solving the three long-standing pain points of the VC industry: "data dispersion, network islands, and manual research." The Scout AI agent's intent understanding ability and native integration of the MCP protocol make it one step ahead in "AI-native" attributes compared to traditional competing products. US$30 million in financing and dual SOC 2/ISO 27001 certifications demonstrate that it has established a foundation of trust in the professional institutional market.

Current limitations: The product only supports English, and there is no clear plan for the coverage of the Chinese market and Chinese companies; there is no public pricing and personal packages, and the purchase threshold is high; as a younger platform, the depth of historical data (such as going back more than 5 years) and late-stage company coverage may not be as good as those of older databases such as PitchBook.

Adoption risk assessment: Before purchasing, you need to confirm whether the data dimension coverage meets the team's industry and regional needs; the rate limit and batch query cost of the API solution need to be verified through PoC; for teams that are deeply bound to PitchBook or Crunchbase, switching costs (data migration, workflow reconstruction, employee training) may exceed expectations. It is recommended to apply for a Demo trial in the name of a single fund or team first, and then use actual deal flow scenarios to verify the data quality and Scout's accuracy before deciding whether to extend it to the entire organization.

Version Info

• Scout AI Agent + MCP integrated version ：There is no official precise date yet; this version launches the AI ​​intelligent agent Scout, supports natural language market research MCP protocol integration with Zapier/n8n automated workflow, and the Harmonic Hot 25 Startups quarterly list. （~2026-06）
• API + network mapping version ：There is no official precise date yet; REST API, Chrome extension LinkedIn/calendar network mapping function are opened, and Bulk Data data warehouse delivery is added. （~2025）
• Search and Monitor Edition ：There is no official precise date yet; advanced search and filtering, customized monitoring and alarming, and team collaboration workbench are launched to allow enterprise customers to access it on a large scale. （~2024）
• Initial database version ：There is no official precise date yet; the first public version provides a startup company database and basic search functions, and is in the seed user verification stage. （~2023）