Glasser

Firecrawl vs Exa: choose the right API for tasks that need web data

What Firecrawl and Exa can do

Limits & data comparison

Compare by
Firecrawl
Direct API
Exa
Direct API
Limits

Search & scrape: 100 requests / min

Hobby: 20 crawl submissions / min; 5 concurrent browsers

Search: 10 / sec; Contents: 100 / sec

Default team limits; deep search modes: 5 / sec

Data volume

10,000 pages / crawl by default

Set a lower crawl limit to control page usage

Up to 100 URLs / Contents request

Search: up to 100 results, depending on search type

Data fields

Title, description, language, source URL and page status

Optional screenshots, links and schema-based JSON

Title, URL, author and publication date when available

Optional summaries, links and image links

Collection & context controls

Crawl depth, URL paths and HTML tag filters

Main-content extraction and query-guided highlights

Domain and publication-date filters

Text-length limits, query-guided highlights and optional subpages

Execution & Reliability

Synchronous scrape; asynchronous crawl with polling or webhooks

Scrape timeout: 60s default, adjustable from 1–300s. Inspect crawl status and the Crawl Errors endpoint for reported page failures and robots.txt blocks.

Search & Contents return results in the same request

Contents live-crawl timeout: 10s default, adjustable up to 90s. Check per-URL statuses and error details to identify failed fetches.

Freshness

Default cache window: 48 hours

maxAge: 0 fetches fresh; minAge: 1 uses cache only (404 if absent)

Cached content first; fetch live if absent

maxAgeHours: 0 fetches fresh; 1–720 sets cache age; −1 uses cache only

Firecrawl vs Exa pricing

Firecrawl

$19/ month

Hobby · billed monthly

Included
5,000 credits / month
Basic scrape or crawl
1 credit / page
Free tier
1,000 credits / month
Annual billing
$16 / month, billed annually

Exa

$7/ 1,000 searches

Auto Search

Included
Up to 10 results / search
Contents
$1 / 1,000 pages per content type
Free tier
$10 credit / month
Signup bonus
$10 credit

Monthly cost example

3,000 web searches + 2,000 page extractions

Monthly usageDirect APIsThrough Glasser
3,000 Exa Auto searches$21.003,000 × $0.007$21.003,000 × $0.007
2,000 Firecrawl page extractions$19.00Hobby monthly plan · 5,000 credits$2.142,000 × $0.001071
Monthly total$40.00$23.14

Inputs and outputs

Firecrawl

  1. InputA public page URL
  2. OperationScrape the page
  3. OutputMarkdown and HTML

Exa

  1. InputA research question
  2. OperationSearch and request highlights
  3. OutputLinks and relevant passages

Still unsure? Choose based on your task

Why Use Firecrawl and Exa Through Glasser?

One API Key

Access both providers with one Glasser key.

One Balance

Use one shared balance for both providers.

Pay as You Go

Pay per use, with no Glasser subscription.

Agent-Ready Integration

Connect your agent via Skill, MCP or CLI.

Questions before you choose

Can I use Firecrawl and Exa together in one AI agent?

Yes. Your agent can use Exa to find relevant URLs, then Firecrawl to capture selected pages. Glasser provides access to supported endpoints under one account and shared balance, without separate provider accounts. Your agent coordinates the calls.

Can my agent automatically choose between Firecrawl and Exa?

A connected agent that supports tool use can search Glasser’s endpoints, inspect their inputs and prices, and select an operation for your task. The agent makes that choice based on your instructions and the endpoints it can access. You can specify a provider or require approval before paid calls.

Can I use Glasser with Claude Code, Codex and ChatGPT?

Yes. For local Claude Code and Codex tasks, Glasser recommends the Skill plus CLI; MCP is also supported. ChatGPT connects through Glasser’s MCP integration, with no local CLI installation. Follow the setup guide for your client and check your balance to confirm the connection for free.

Does Glasser provide every Firecrawl or Exa feature?

The endpoints on Glasser expose specific operations and supported inputs. For example, the basic Firecrawl scrape endpoint here supports Markdown and HTML capture, with no browser actions, PDF parsing or AI formats. Inspect the selected endpoint before relying on a provider feature.

Is using Glasser always cheaper than going direct?

It depends on your workload. Small workloads can fit the providers’ direct free tiers, and larger workloads can favor a subscription. The example above shows savings for its stated volumes and credit assumptions. Compare the same operations, result counts and billing period, including any free credits you can use.

How do I control API spending through Glasser?

Search and Inspect are free. Ask your agent to show the price and planned volume, then wait for approval before paid calls. Start with small batches and review each Run’s final charge; unsuccessful Runs can still be chargeable under the endpoint’s rules. Your balance, including signup credit, is shared across providers.

Get the data you need with Glasser.