Exa

Web search rebuilt for LLMs

exa.ai

HQ: San Francisco, CA, USA

11-50

B2B

Company research

Public company, workplace, funding, and market signals

Updated Aug 2, 2026

Overview

Exa (formerly Metaphor) is a San Francisco applied-AI research lab building web search and data infrastructure for AI agents. It crawls and indexes the web, trains neural retrieval models, and exposes search, crawling, contents, monitoring, and research-agent capabilities through APIs and products.

Primary product

Exa Search: an AI-native web search API for agents and applications, complemented by Contents, Websets/Deep research, Monitors, and Exa Agent.

Founded

2021

Headquarters

San Francisco, California, United States

Team size

75-100 (public signals range from YC's 75 to LinkedIn's 97; time-sensitive)

Work style

Onsite

Industry

Software Development

Sub-industry

AI search, web data infrastructure, retrieval and developer APIs

Offices

San Francisco, California, United States

Singapore (engineering hub; publicly announced hiring location)

Business model

Pay-as-you-go API usage

Enterprise contracts with volume discounts, SLAs, custom indexes and Zero Data Retention

Developer free tier that converts to usage-based API revenue

Funding

Stage

Series C / venture-backed private company

Total raised

$361.4M

Latest round

Series C · May 2026

Latest amount

$250M

Series C

May 2026 · Andreessen Horowitz (a16z)

$250M

Series B

Sep 2025 · Benchmark

$85M

Series A

Jul 2024 · Lightspeed Venture Partners

$17M

Seed

Jan 2022

$5M

Pre-seed

Jan 2021

People & investors

Leadership

Will Bryk

Co-founder and CEO

Jeffrey (Jeff) Wang

Co-founder

Marcus Holm

Chief Revenue Officer

Jeeyoung Kim

Head of Engineering

Founders

Jeffrey Wang

Co-Founder

Investors

Andreessen Horowitz

Andreessen Horowitz (a16z)

Benchmark

Igor Babuschkin

Lightspeed Venture Partners

Work & culture

The company describes itself as a worldwide team of builders, engineers and researchers working on technically difficult web-scale search. Public recruiting signals emphasize in-person collaboration, ML reading groups, lunch talks and high ownership.

Work style

Onsite

Visa sponsorship

Yes

Compensation

Public job postings and recruiting materials indicate competitive startup compensation, but no reliable company-wide salary or equity bands were found in the reviewed sources. Benefits and visa sponsorship are explicitly documented.

Benefits

Medical, dental and vision coverage

Monthly location-tiered wellness benefit

SF commuter benefit

Fertility and family-planning reimbursement, including up to $15,000 lifetime for eligible egg-freezing expenses

16 weeks fully paid parental leave for all new parents

Daily lunch and dinner

Snacks, coffee and health supplements

Product & market

Pricing

Usage-based pay-as-you-go: public pricing lists Search at $7 per 1,000 requests, Contents at $1 per 1,000 pages, Answer at $5 per 1,000 requests, Monitors at $15 per 1,000 requests, and Agent at approximately $0.012-$1.00 per run depending on effort; enterprise pricing is customized.

Differentiators

—Designed specifically for AI agents rather than human click behavior
—Neural/embedding retrieval and link-prediction models rather than only keyword ranking
—First-party web index, crawler and custom vector database rather than simply wrapping another search engine
—Low-latency and high-compute search modes
—Structured outputs, citations, page contents and extensive filtering/customization
—Zero Data Retention and enterprise controls
—Scale claims include hundreds of billions of tracked URLs and hundreds of thousands of searches per second

Technology

Web crawling and distributed crawling infrastructure

Specialized transformer retrieval models

Embedding-based semantic search

Custom web-scale vector database

Matryoshka embeddings

Binary quantization

C++ performance optimization

AWS

Kubernetes

PyTorch

GPU clusters including H200s

Custom HTML parsing and metadata filtering

Customers

Cursor

Cognition

HubSpot

OpenRouter

Monday.com

Databricks

More than 5,000 corporate customers claimed in the May 2026 Series C announcement

Competitors

Google Search

Perplexity

Brave Search

Bing

Tavily

SerpAPI and other search/API wrappers

Scale & contact

Estimated revenue

Approximately $10M-$12M annualized revenue/ARR as of late 2025, based on third-party estimates; not company-reported and should be treated as directional.

Estimated monthly visits

465.7K

support@product.exa.ai

Latest news

Introducing Exa Agent

Exa Blog · Jun 2026

Product launch for Exa Agent, positioned as frontier web research with strong price/performance and integrated with Exa’s search stack for agents.

Andreessen-Backed AI Search Startup Exa Valued at $2.2 Billion

Bloomberg · May 2026

Independent coverage of the Series C and valuation, describing Exa's positioning against established web search.

Exa Raises $250M Series C to Build the Search Engine for AIs

Exa Blog · May 2026

Exa announced a $250M Series C led by a16z at a $2.2B valuation, citing more than 5,000 companies using the platform and plans to scale models, infrastructure and go-to-market.

Exa Raises $85M to Build the Search Engine for AIs

Exa Blog · Sep 2025

Exa announced an $85M Series B led by Benchmark at a $700M valuation; Peter Fenton joined the board.

How We're Building the Next Generation of Search with Semantic Search Technology

Exa Research Blog · Mar 2025

Technical explanation of Exa's crawler, embedding-based retrieval, custom vector database, Search API and Websets architecture.

Jobs and careers at Exa

52 jobs at Exa

IT Engineer

On-site

Engineering

San Francisco, California; San Francisco, California, United States

3d ago

On-site

Full-time