Exa Agent Brings $2.2B Semantic Search Directly Into Google Sheets

Exa, the semantic search API company that recently hit a $2.2B valuation, has launched its Exa Agent directly inside Google Sheets. The move takes a capability that was previously only accessible via API calls and puts it in front of anyone who can open a spreadsheet. If you have a column of company names and want to fill in funding rounds, executive names, or competitor lists, you can now do that with a sidebar click.

What Exa Actually Is

Exa is a web search API designed for teams building applications that need real-time web data. It performs semantic search using embeddings, extracts clean content from search results, and provides direct answers with citations. The key difference from a standard search API is how it finds things: unlike traditional keyword search APIs that return pages matching exact keywords, Exa uses neural embeddings to find content semantically similar to your query, retrieving pages that express the same ideas or information even without exact word matches.

Today, Exa claims to be the highest quality search API at every latency, powering search for Cursor, Cognition, HubSpot, OpenRouter, Monday.com, and over 400,000 developers. The Google Sheets add-on is a direct attempt to bring that infrastructure to non-developer users.

The Exa Agent: What Changed

The headline feature in this Sheets update is Exa Agent, which Exa launched just days before the Sheets announcement. This is not just a search function. Exa Agent combines top language models with Exa's state-of-the-art web search tools to achieve exhaustive and accurate results, and it is highly effective on deep research, list-building, and entity enrichment tasks.

When you throw a large dataset at it, Exa Agent divides the task into many subtasks and assigns subagents to research various domains at once, using a fusion of frontier and cost-effective models to find the most cost-effective methodology for the given research task. The result is a system that can work through hundreds of rows in parallel, something that would take hours manually.

The Full Function Toolkit

Beyond Exa Agent, the Google Sheets add-on ships with five spreadsheet functions that cover most research and enrichment workflows:

• =EXA(prompt): The simplest entry point. Describe what you want about a cell value and it returns it. For example: =EXA("Return only the CEO name for " & A2)
• =EXA_ANSWER(prompt, prefix, suffix, includeCitations): Generates an AI answer from a deep web search, with optional citations and structured output.
• =EXA_SEARCH(query, numResults, searchType): Returns an array of URLs. Supports three search modes: "fast", "auto", and "deep".
• =EXA_CONTENTS(url): Extracts the full text of any webpage directly into a cell.
• =EXA_FINDSIMILAR(url): Finds URLs semantically similar to a given page, useful for competitor discovery.

The sidebar also includes a batch refresh feature to update multiple cells at once instead of manually editing each formula. There is also a "Convert to Values" option to freeze results and prevent auto-refresh charges from accumulating as your sheet recalculates.

Dynamic Queries: Where It Gets Interesting

The real power comes from combining Exa functions with standard spreadsheet logic. You can build queries dynamically from cell values using the & operator, which means a single formula dragged down a column can run a unique web search for every row. A competitor analysis sheet, for example, might look like this:

=EXA_SEARCH(A2 & " " & B2 & " market analysis", 5, "deep")

Where column A holds company names and column B holds their industry. Each row fires a separate, targeted search. You can also combine Exa with the Claude for Sheets add-on: Exa brings the power of semantic search directly into Google Sheets, letting you query the web, retrieve relevant results, and enrich your spreadsheets with up-to-date information, all without leaving your worksheet. Claude can then process and summarize what Exa retrieves.

Pricing and What It Costs to Run

The add-on itself is free to install, but every function call hits the Exa API, which has usage-based pricing. Be mindful of API rate limits when running large batch operations, as each function call counts as one API request. For Exa Agent specifically, pricing is tiered by effort level:

Effort	Cost per Request	Best For
minimal	$0.012	Lightweight tasks, lowest cost
low	$0.025	Simple lookups, narrow factual tasks
medium	$0.10	Default for most standard research
high	$0.50	Harder research, more citations
xhigh	$1.00	High-value tasks where completeness matters

The auto effort mode dynamically scales compute to the task, which is the recommended starting point. For a sheet with 100 rows of company enrichment at medium effort, you are looking at roughly $10 per full run, which is competitive with manual research tools or dedicated data enrichment platforms.

Who This Is Actually For

This release sits at an interesting intersection. Developers building AI agents, researchers conducting market analysis, and sales teams sourcing leads can use Exa to access structured web data. The Sheets integration specifically targets the non-developer end of that spectrum: growth teams building prospect lists, analysts enriching CRM exports, and researchers who want live web data without spinning up a Python environment.

Exa is backed by Lightspeed, NVIDIA, and Y Combinator, and recently raised a $250M Series C. The Google Sheets add-on is free to install from the Google Workspace Marketplace. To get started, install the add-on, open the sidebar via Extensions > Exa AI > Open Sidebar, and paste in your API key from dashboard.exa.ai.