Editor's Verdict

The Data Layer Under AI Visibility Tools,
Sold Straight to You

4.2
★★★★☆
Very Good
Most AI visibility tools sell you a dashboard. Cloro sells you the pipe underneath it. One request, one JSON shape, and back comes what ChatGPT or Google's AI Mode said, which pages it cited, and the searches it ran to get there. The free tier is real (500 credits a month, no card), and you only pay for calls that succeed. The trade-off shows up the moment you log in: there is nothing to look at. If you can't write a script or wire up n8n, this isn't your tool.
Features & pricing last verified: September 2026

What We Love

  • Nine data sources behind one schema, including the AI engines SERP APIs skip
  • Sources, citation pills and query fan-out come back parsed, not buried in HTML
  • Failed requests cost nothing, and retries are automatic
  • Free tier, published credit prices, and a changelog that ships almost weekly

! Could Be Better

  • API only. No reporting, no alerts, no charts
  • 30 to 45 seconds a request; fine for monitoring, useless for anything live
  • Credits expire monthly, and the AI platforms keep changing what they return
Not for you if: You want to log in and read a report. Rankscale, Rank Prompt and OmniSEO give you brand mentions and share of voice without writing a line of code. Cloro is for the people who build those views themselves.
✓ Free tier, no card • Paid plans from $30/mo Try Cloro →

What Is Cloro?

A scraper you never have to maintain, for the search engines that talk back.

Cloro (cloro.dev) is a developer API for search and AI-answer data. You send it a prompt and a country code, and it opens a real browser session against the live interface of whichever platform you asked for (ChatGPT, Perplexity, Microsoft Copilot, Gemini, Grok, Google Search, Google News, Google's AI Overview or Google's AI Mode), reads what a person in that country would have seen, and hands it back as JSON. The answer text, the sources, the inline citation chips, the search queries the AI fired off behind the scenes, and, depending on the endpoint, shopping cards, local packs, ads and knowledge panels. Same shape every time, whichever engine you asked.

That consistency is the point. Anyone who has tried to scrape ChatGPT logged-out, or parse a Google page that changes its class names every few weeks, knows the maintenance never ends. Cloro's pitch is that you stop doing it. The company launched the API in October 2025 and has shipped a changelog entry most weeks since: new endpoints, new fields, new countries, and a steady stream of fixes for things the platforms broke. Reading that changelog is the best way to understand what you're buying. The product is a moving target, and the team seems to know it.

Where it sits in the market matters. The AI visibility category (tracking whether ChatGPT or AI Overviews mention your brand, and who they cite instead) has split into two layers. On top are the dashboards: Rankscale, Rank Prompt, OmniSEO, the AI visibility modules inside Semrush and SE Ranking. Underneath is the plumbing that fetches the answers, and several of the dashboards run on APIs like Cloro's. Cloro sells that plumbing directly. It has no opinion about how you should chart it.

How We Evaluated Cloro

I signed up for the free tier, created an API key, and ran a synchronous ChatGPT request from my terminal asking what ChatGPT knows about Max Productive AI. Then I went through the dashboard, the billing page and the playground, read the public changelog back to launch, and checked every price against the docs. What I say about latency and reliability comes from public user reviews and independent write-ups, not from a load test. The performance numbers Cloro publishes (99.9% uptime, a 98% AI Overview trigger rate on commercial queries) are its own claims. Nobody has run a neutral head-to-head benchmark of these APIs yet, and I haven't either.

So who is it for? Three groups. SEO and GEO teams with a developer who want to track hundreds of prompts across every engine and keep the raw history in their own database. Agencies and SaaS companies building an AI visibility feature into their own product. And technical marketers who are comfortable in n8n or a Python notebook and would rather pay $30 a month for data than $99 or more for someone else's charts. If none of that sounds like you, the dashboards further down this page are the better read.

See Cloro in Action

Screenshots from our free account, from the first API call through to the billing page.

1

A Real ChatGPT Request, End to End

One curl command, one JSON response, with the citations broken out

Terminal showing a curl POST to api.cloro.dev/v1/monitor/chatgpt and the JSON response with text, sources and citationPills fields
EndpointPOST /v1/monitor/chatgpt with prompt + country
Parsed Fieldstext, sources, citationPills
AttributionEvery URL tagged utm_source=chatgpt.com

If you only look at one screenshot, make it this one. I asked ChatGPT, through Cloro, what Max Productive AI is and what it offers. Twenty seconds or so later I had the full answer as text (and it was accurate, right down to our affiliate disclosure, which was a slightly odd thing to read back), a sources array with our homepage flagged as a footnote source, and a citationPills array listing every page ChatGPT linked inline: the homepage, the productivity category, the email management category. Notice the utm_source=chatgpt.com on each URL. That is how ChatGPT tags its outbound links, and Cloro leaves it intact, handy when you're matching this data against what shows up in your analytics. Also notice "model": null. That's not a bug on Cloro's side. OpenAI moved almost all logged-out traffic to a new interface format in late August 2026 that no longer exposes the model name, and Cloro's changelog says so plainly. More on that below.

2

API Key Management

Create a key, keep it private, and note the bonus-credit offers in the sidebar

Cloro API Keys page with one active sk_live key created Sep 8, 2026, and a Get started sidebar offering bonus credits
Keyssk_live_ prefix, created and revoked in one click
Bonus Credits+500 for a review, +500 for a backlink, +200 on Reddit
Upgrade BannerPersistent on every dashboard page

Nothing surprising here, which is the point. One key per organisation to start, a created date, an active flag, a bin icon. The bit I'd draw your eye to is the "Get started" box bottom left. Cloro will top up your free tier by 500 credits a month for leaving a public review, another 500 for linking back to cloro.dev, and 200 for saying hello in its community. That's a growth tactic, and a fairly transparent one. Do the maths though: 1,200 extra credits a month roughly triples the free tier, which for a solo developer prototyping something is a decent deal. Just be aware that any review you read of Cloro elsewhere may have been written with that incentive in view. This one wasn't.

3

Dashboard: Credits and Usage

Remaining, used, and a projection for the month, per day

Cloro dashboard showing 486 credits remaining, 14 credits used, a projected 55 credits for the month, and a daily usage bar chart
Balance486 of 500 credits left after testing
ProjectionEstimates month-end spend from your run rate
FiltersCredits or requests, by day, week or month

My test calls on the 8th cost 14 credits between them, so the dashboard shows 486 remaining and projects 55 for the month. This screen is the only reporting Cloro gives you, and it's reporting about Cloro, not about your brand. Credits, requests, a daily bar chart, and a Logs page with each request's status. That's fine for what it is. It just needs saying clearly, because "dashboard" means something different in this category, and if you were hoping to see which pages ChatGPT cites for your keywords, you'll be building that yourself.

Want to see what ChatGPT says about your brand?

Try Cloro →500 free credits a month • No card required
4

Playground: Nine Endpoints

Three steps to a first request, and the full list of what you can query

Cloro playground with three onboarding steps and endpoint links for ChatGPT, Copilot, Perplexity, Grok, Gemini, Google Search, Google News, Google AI Mode and Google AI Overview
AI EnginesChatGPT, Copilot, Perplexity, Grok, Gemini
GoogleSearch, News, AI Mode, AI Overview
Try ItInteractive panel lives in the docs, not here

"Playground" oversells it slightly. It's a three-step guide: create a key, pick an endpoint, and use the Try It panel on that endpoint's documentation page. Each of the nine links opens the API reference in a new tab. I'd have liked to fire a request from inside the dashboard, but the docs route works, and the docs themselves are good: full schemas per endpoint, worked examples, and an AI assistant that answers questions with citations back to the relevant page. What the screen does well is show the breadth at a glance. Five AI engines and four Google surfaces from one account is more than any SERP API I know of. That breadth is the main reason to sign up.

5

Pricing Plans

Free through Enterprise, all on one page, with per-credit rates spelled out

Cloro billing page showing Free, Lite $30, Hobby $100, Starter $250, Growth $500, Business $1,000 and Enterprise plans with credits and concurrent jobs
EntryLite at $30 for 37,500 credits
Sweet SpotHobby at $100 halves the per-credit rate
Concurrency1 job free, 10 on Lite, 100 on Business

Every tier on one screen, with credits, price per thousand credits and concurrent jobs listed for each. I appreciate that; too many API vendors hide the mid-tier behind a sales call. Look closely at two numbers, though. The per-credit rate drops sharply from Lite ($0.80 per 1,000) to Hobby ($0.40) and then flattens, so Hobby is where most serious users should start. And the number that really gates throughput is concurrent jobs, not credits: one job at a time on the free tier means your 500 credits drip out slowly whether you like it or not. The "Popular" badge sits on Business at $1,000. I suspect that says more about who Cloro's customers are (other tools) than about where a small team should land.

How Cloro Works

From a prompt to a row in your database, in four steps.

1

Create a Key and Pick an Endpoint

Sign up, generate an sk_live_ key, and choose the platform you want to read. Each has its own URL under /v1/monitor/: chatgpt, perplexity, copilot, gemini, grok, aimode, google and google/news. Google's AI Overview isn't a separate endpoint; you ask for it with include.aioverview on a Google Search request. Country availability differs per platform (ChatGPT won't serve logged-out traffic from the Czech Republic, for example), so the /v1/countries?model= endpoint is worth hitting before you deploy.

2

Send a Prompt with a Country

The minimum request is a prompt and a two-letter country code. From there you layer on options. hl sets the interface language independently of the country on the Google endpoints. state routes AI requests through a specific US state for +2 credits. location or uule pins Google Search and AI Mode to a city. include.markdown gives you the answer with inline links; include.searchQueries, include.shopping or include.ads on ChatGPT add +2 credits between them; disableWebSearch shows you what ChatGPT says when it isn't forced to search. Then decide: synchronous, and wait 30 to 45 seconds for the answer in the same call (+2 credits), or asynchronous, and get a task ID back instantly.

3

Read the Parsed JSON

Every endpoint returns result.text, a sources array and, where the interface shows citation chips, a citationPills array with a label, URL, domain and position for each source, grouped by chip. ChatGPT, Copilot and Perplexity add searchQueries, the fan-out queries the AI ran before answering, which is the closest thing this industry has to seeing the AI's keyword research. Beyond that it depends on the surface: shopping cards and inline products on ChatGPT, Copilot, Perplexity and AI Mode; map entries and local packs; ads inside AI Overviews and ChatGPT; a Knowledge Graph object on Google Search. The raw HTML is available on request and expires after 24 hours.

4

Run It on a Schedule, Keep the History

Monitoring means repetition, and the async endpoints are built for it. Submit up to 500 tasks in one batch call, give the urgent ones a priority from 1 to 10, and receive results by webhook (signed with HMAC so you can verify they came from Cloro) or by polling. Completed tasks are kept for 24 hours only, so your job is to write each result somewhere permanent: Postgres, BigQuery, a Google Sheet via n8n, whatever you already have. Cloro stores nothing for you beyond that window. The dashboard shows credits and logs, not trends. The trends are yours to build.

Sync or Async?

Sync is for testing and one-offs. It costs +2 credits, waits up to five minutes with as many as ten automatic retries, and if the request still fails you're not charged. Async is for everything else: no surcharge, batch submission, webhooks, and queued tasks aren't billed until they run. If you hit a backlog, DELETE /v1/async/queue clears it. Rate limits are 1,000 requests a second per endpoint per key, which nobody on a self-serve plan will touch. Concurrency is the real ceiling.

The Ground Moves Under You

Cloro reads the same interfaces a human sees, so when a platform changes, the data changes. The changelog is candid about this. On 25 August 2026 OpenAI finished rolling out a new mobile-web format that serves 98% of ChatGPT responses and carries no model name or search-query metadata, so result.model and result.searchQueries now come back empty most of the time (a legacy: true flag is offered as a best-effort workaround). Perplexity tightened access the same month and the endpoint went from 3 to 4 credits. Google started serving encrypted /goto redirect links, which Cloro now resolves for you. None of this is Cloro's fault, and the response has been quick each time. But build your parser to tolerate empty fields, and read single data points as noise. The value is in the trend.

Key Features

What's in the response, and what costs extra credits.

Core

Nine Sources, One Schema

ChatGPT, Perplexity, Copilot, Gemini, Grok, Google Search, Google News, AI Overview and AI Mode, with text, sources and citation pills in the same shape on each. Write the parser once. DataForSEO, SerpApi and Serper.dev don't offer native ChatGPT, Perplexity or Copilot endpoints at all. Cloro exists to fill that gap.

Core

Citation Pills and Sources

The inline chips an AI shows next to a claim, broken out per source with label, URL, domain and position, and grouped by chip so you can see which pages share a citation. ChatGPT adds a publish date. This is the field AI visibility tools are built on, and getting it as JSON rather than regex-ing it out of markdown is the difference between an afternoon's work and a week's.

Add-on on ChatGPT

Query Fan-Out

The searches the AI ran before it answered. Copilot and Perplexity return them always; ChatGPT needs include.searchQueries (+2 credits) and, since the August interface change, often returns them empty unless you try the legacy flag. When it works it's gold: it tells you the actual keywords behind an AI answer, which is what you should be ranking for.

Core

Geo-Targeting Down to a City

250 countries for Google Search, around 195 for the AI engines, and 211 for AI Mode. US state targeting for the AI endpoints costs +2 credits; city-level location or a raw uule string on Google Search and AI Mode costs nothing. A separate hl parameter sets the interface language, so you can query Google in German from a US location, or French from Switzerland, without faking the country.

Mixed

Shopping, Ads and Local Packs

Product cards and inline products from ChatGPT (include.shopping, +2), Copilot, Perplexity and AI Mode; sponsored ads inside AI Overviews, ChatGPT and Google Search, with a flag on ChatGPT ads telling you which one was actually rendered; map entries with ratings and place IDs; Google's local pack and Knowledge Graph panel. For an e-commerce or local brand this is the difference between "are we mentioned" and "are we the product card".

Core

Async, Batch and Webhooks

Up to 500 tasks per batch call, priorities from 1 to 10, idempotency keys, a queue-status endpoint, and signed webhook deliveries with copy-paste verification code for Node, Python and Go. Each response carries an X-Latency-Ms header. This is the part of Cloro that feels most like it was built by people who've run monitoring pipelines themselves.

Core

MCP Server and Integrations

A hosted MCP server at mcp.cloro.dev turns every endpoint into a tool for Claude, Claude Code, Cursor or your own agent, so you can ask "does Google's AI Overview mention us for X?" in a chat and get a real answer. There are also n8n, Zapier, LangChain and OpenClaw integrations. For non-developers, n8n plus a Google Sheet is the quickest route to something useful.

Core

Success-Only Billing

Failed requests aren't charged. Sync calls retry up to ten times inside a five-minute window on Cloro's side, so you can drop your own retry logic. Queued async tasks don't bill until they run. GET /v1/credits returns your balance and reset date for low-balance alerts. The one time an incident affected paid output (a 47-minute HTML outage in August), the changelog told customers to ask for a refund of the affected credits.

A couple of caveats before you get carried away. First, a lot of the richest fields are opt-in and cost extra: shopping, ads, raw responses and search queries on ChatGPT all sit behind a +2 credit flag, and US state targeting is another +2. Budget from the credit table in the next section, not from the base price. Second, Grok sits in the list and in the playground, but it's had the roughest ride of any endpoint (a price increase in May over anti-bot defences), so check the docs for its current status before you build on it.

See the full endpoint list and credit prices:

Check Cloro's pricing →Free tier • Lite from $30/mo • Hobby from $100/mo

Cloro Pricing

Credits, not calls. Work out what a request costs before you pick a plan.

Free

$0/mo
  • ✓ 500 credits, refreshed monthly
  • ✓ Every endpoint, every country
  • ✓ Async, webhooks, MCP server
  • ✓ No card required
  • ✓ Bonus credits for review, backlink, community
  • ✗ 1 concurrent job only
Start Free

Lite

$30/mo
  • ✓ 37,500 credits/month
  • ✓ $0.80 per 1,000 credits
  • ✓ 10 concurrent jobs
  • ✓ Roughly 5,000 to 7,000 AI answers/month
  • ✓ Everything in Free
  • ✗ Credits expire each cycle
View Lite Plan
Higher tiers: Starter $250/mo (650,000 credits, 50 jobs), Growth $500/mo (1.35M credits, 75 jobs), Business $1,000/mo (2.8M credits, 100 jobs), Enterprise $2,000 to $5,000/mo (5.9M to 16.4M credits, up to 255 jobs)
Enterprise custom: Above $5,000/mo, with higher concurrency, credit overages and annual discounts. Contact Cloro
Rollover: None. Unused credits expire at the end of each monthly cycle
Pricing last verified September 2026. Visit Cloro for current rates.

What a Request Costs

EndpointBase creditsCommon add-ons
Google Search (1 page)3+2 AI Overview, +2 per extra page (up to 10)
Google News (1 page)3+2 per extra page
Google AI Mode4+1 per product cluster with expandProducts (max +6)
Perplexity4Was 3 until August 2026
Gemini4None
Grok4None
Copilot5None
ChatGPT~5 to 7+2 for shopping, ads, rawResponse or searchQueries (any combination)
Any AI endpoint+2 US state targeting, +2 synchronous call

Is It Worth the Money?

250,000
Hobby credits per month
=
~$0.003
per AI answer, sources included

Say you track 100 prompts, daily, across five AI engines. That's 15,000 requests a month, or somewhere between 60,000 and 100,000 credits with a few add-ons. It fits inside Hobby with room to spare, for $100. The dashboard tools charge a similar amount for a fixed and much smaller prompt allowance, and you don't get the raw data back. The catch is the labour: that $100 buys data, and someone still has to turn it into something a client can read.

The free tier deserves a fair word. 500 credits with no card is enough for around 70 sync ChatGPT calls or 100-odd Google pages a month, which is enough to build a prototype on, not just poke around. The constraint isn't credits, it's the single concurrent job: everything runs one at a time. If you're a solo developer testing an idea, add the bonus credits and you can live on it for a while. Once you're monitoring anything for real, Lite's ten jobs matter more than its 37,500 credits.

Before you commit, get your head around two quirks of the billing. Credits don't roll over, so an occasional deep audit is better done on the free tier plus a one-month upgrade than on a standing plan. And depth is priced honestly but not cheaply: since May 2026, every extra Google page costs +2 credits, so a 10-page pull is 21 credits rather than the 5 it briefly was. Cloro explains why (deep pagination fails more and costs it more), and I'd rather have that than a headline price that quietly caps you. Against the SERP-only vendors, Cloro is more expensive per plain Google page than Serper.dev or DataForSEO's standard queue and cheaper than SerpApi at depth. But nobody buys Cloro for plain Google pages. You buy it for the AI endpoints, and there the comparison is with dashboard tools, where it wins on price and loses on convenience.

Detailed Pros & Cons

The full picture, including the parts you only learn from the changelog.

✓ Pros

The Broadest AI-Engine Coverage in One API

ChatGPT, Perplexity, Copilot, Gemini and Grok alongside Google Search, News, AI Overview and AI Mode. The established SERP APIs don't touch the chat engines, and the dashboard tools don't hand you the raw responses. Cloro is the only product I've found that does both. If you're building AI visibility tracking, that settles it.

Output You Can Use Without Parsing

Text, sources, citation pills with chip grouping, query fan-out, shopping cards, map entries, ads, Knowledge Graph. My ChatGPT test came back with every cited URL already separated out and tagged. Public user reviews land on the same point again and again: the win isn't the scraping, it's not having to write and maintain the parser.

You Only Pay for What Works

Failed requests are free, retries are automatic (up to ten, inside five minutes), and queued async tasks aren't billed until they run. When an August incident broke HTML links for 47 minutes, the changelog told customers to claim the affected credits back. That's the right instinct, and it's rarer than it should be.

A Free Tier That Isn't a Trick

500 credits a month, no card, all endpoints, async and MCP included. Bonus credits can roughly triple it. You can build and test a working prototype without paying anything, which is more than most API vendors offer.

Serious Pipeline Tooling

Batch creation of 500 tasks, task priority, signed webhooks, queue status and queue clearing, a credits endpoint, per-request latency headers, per-endpoint rate limits. These are the features you only notice when they're missing, usually at 2am.

A Changelog That Tells You the Truth

Entries most weeks since October 2025, and they include the bad news: fields that went empty because OpenAI changed formats, a Perplexity price rise with the cost reasoning attached, countries removed because a platform put up a login wall. Read it before you buy. It's the most useful document Cloro publishes.

✗ Cons

No Reporting Layer at All

The dashboard shows credits, keys and logs. It does not show brand mentions, share of voice, cited domains or anything about the data you fetched. You will build that yourself, or pipe the JSON into n8n, a sheet or a warehouse. For a marketing team without a developer, that's a dealbreaker, and Cloro doesn't pretend otherwise.

Slow by Design

Thirty to forty-five seconds a request, and around twenty for ChatGPT, because a real browser session is doing real work. Sync calls can wait up to five minutes through retries. Perfect for a nightly batch. Unusable for anything a person is waiting on, which rules out live features in your own product unless you cache aggressively.

The Data Shifts Under Your Feet

In one month of 2026, ChatGPT stopped exposing model names and search queries for 98% of responses, Perplexity got pricier, and Google started encrypting result links. Cloro handled each within days, but a wrapper around live consumer interfaces inherits every change those interfaces make. Expect empty fields, expect breaking changes, and don't build anything that assumes today's schema is permanent.

Credits Expire Monthly

No rollover on any plan. If your usage is bursty (big audit in January, nothing in February), you'll pay for credits you never used. Plan for a steady monitoring cadence, or drop to the free tier between projects.

Add-Ons Add Up

The base credit prices look small until you stack them. A sync ChatGPT call with shopping data from Texas is base plus 2 plus 2 plus 2. Ten pages of Google with AI Overviews is 23 credits, not 3. None of it is hidden, but the request you need can easily cost three or four times the headline figure.

Performance Claims Are Self-Reported

The 99.9% uptime, the 98% AI Overview trigger rate on commercial queries, the favourable reliability comparisons against DataForSEO and SerpApi: all of it comes from Cloro's own blog and docs. It may well be accurate. But no independent lab has benchmarked these APIs side by side, so treat the numbers as marketing until someone does.

Cloro Alternatives: How It Compares

Two kinds of rival: other data APIs, and the dashboards that could be built on them.

FeatureReviewed CloroDataForSEOSerpApiSerper.devRankscale
TypeRaw APIRaw APIRaw APIRaw APIDashboard
Entry PriceFree; $30/mo$50 deposit, pay as you go$25/mo (1,000 searches)Free 2,500; $50 packFree; $29/mo
ChatGPT / Perplexity / Copilot✓ Native endpoints✓ Tracked in-app
Google AI Overview✓ Parsed, +2 creditsAdd-on, lighter parsing✓ Parsed✗ No structured field✓ Tracked in-app
Billing at Depth+2 credits per pagePer page, cheapest bulkPer batch of 10 results2 credits above 10 resultsPer prompt allowance
Standout StrengthAI-engine coverage, one schemaBulk cost + backlinks, keywords, on-page80+ engines, legal indemnityFastest, cheapest Google-onlyReady-made reports
Best ForBuilding AI visibility trackingProgrammatic SEO at scaleMulti-engine appsCheap Google dataMarketers without a developer

Cloro vs DataForSEO and SerpApi: The Short Version

If your question is "who cites us in ChatGPT and Perplexity?", the older SERP APIs can't answer it. DataForSEO and SerpApi are excellent at Google, and DataForSEO in particular is cheaper per page and comes with backlink, keyword and on-page data Cloro doesn't have. SerpApi covers 80-plus engines and bundles legal indemnity, which matters to some companies. Neither has native endpoints for the chat engines. So the split is clean: Google-only data at scale, buy DataForSEO. Breadth of engines and legal cover, buy SerpApi. AI answers with citations parsed, buy Cloro. Plenty of teams end up with two of the three. One caution: the most detailed head-to-head comparisons of these vendors online are published by Cloro itself, so check the prices against each vendor's own page.

Build or Buy?

The more common decision is between Cloro and a dashboard. Rankscale, Rank Prompt and OmniSEO track prompts, count mentions, chart share of voice and email you when something changes. They cost about the same per month as Cloro's lower tiers, cover fewer prompts, and don't give you the raw responses. Cloro gives you unlimited flexibility and no help. If you have a developer and want the data in your own systems, or you're building a product, the maths favours Cloro heavily. If you're a marketer who wants answers on Monday morning, it doesn't. Our guide to AI visibility tools compares the dashboards in more depth.

Which Tool Is Right For You?

Rankscale

Dashboard Pick

Best for: Marketers who want AI visibility tracking they can read without a developer. Prompt tracking across the major AI engines, brand and competitor mentions, cited sources and share of voice, all charted for you, from a low monthly price. Less flexible than raw data, but ready on day one.

Rank Prompt

GEO Focus

Best for: Teams focused on Generative Engine Optimization who want to see how AI answers describe their brand and act on it, not just measure it. Rank Prompt pairs monitoring with optimization guidance. If Cloro is the thermometer, Rank Prompt is closer to the treatment plan.

OmniSEO

All-in-One

Best for: Small teams that want AI visibility and classic SEO tracking in one place without stitching tools together. Broader than a pure GEO tracker, simpler than an enterprise suite. You give up the raw API access Cloro offers, and gain a product you can hand to a junior marketer.

Semrush

Enterprise Suite

Best for: Teams already paying for Semrush who want AI visibility as a module next to keywords, backlinks and site audits, rather than a separate tool or API. Far more expensive than Cloro for the AI piece alone, but it's one login, one invoice and a reporting layer your clients already recognise. See how it compares with Search Atlas →

SE Ranking

Agency Value

Best for: Agencies that want AI search tracking bundled into an affordable all-round SEO platform, with white-label reporting. SE Ranking also sells API access to its data, but at a much higher entry point than Cloro and without the chat-engine endpoints. See how it compares with Surfer SEO →

Frequently Asked Questions

Cloro (cloro.dev) is a developer API that returns structured JSON from Google Search, Google News, Google AI Overview, Google AI Mode, ChatGPT, Perplexity, Microsoft Copilot, Gemini and Grok. You send a prompt and a country, and it sends back the answer text, the sources cited, the citation pills and the search queries the AI ran, all in one consistent schema. It is built for SEO and GEO teams and developers who want to build their own AI visibility tracking rather than buy a finished dashboard.
Cloro has a free tier with 500 credits a month, then paid plans from $30 a month (Lite, 37,500 credits) through $100 (Hobby, 250,000), $250 (Starter, 650,000), $500 (Growth, 1.35 million) and $1,000 (Business, 2.8 million). Self-serve Enterprise tiers run $2,000 to $5,000 a month. Each request costs a few credits depending on the endpoint: 3 for a Google Search page, 4 for Perplexity, Gemini, AI Mode or Grok, 5 for Copilot, and about 7 for ChatGPT. Synchronous calls carry a +2 credit surcharge. Credits do not roll over.
Yes, Cloro's free tier gives you 500 credits every month with no card and access to every endpoint, limited to one concurrent job. That is enough for roughly 70 synchronous ChatGPT requests or around 100 Google Search pages a month, which is plenty to test the API and build a prototype. Cloro also offers bonus credits for leaving a public review, linking back to its site or joining its community.
Cloro has endpoints for ChatGPT, Perplexity, Microsoft Copilot, Google Gemini and Grok on the AI side, plus Google Search, Google News, Google AI Overview and Google AI Mode. All of them share the same response shape for text, sources and citation pills, so code written for one endpoint works on the others with little change. Country coverage runs to 250 countries for Google Search and around 195 for the AI endpoints, with some country-specific restrictions.
Cloro has a dashboard for API keys, credits, logs and billing, but not for reporting. There are no brand mention charts, share-of-voice scores or alerts. You are expected to store the JSON yourself and build your own views, or feed the data into a tool like Google Sheets, a data warehouse or n8n. If you want a finished reporting interface, a dashboard product such as Rankscale, Rank Prompt or OmniSEO is the better fit.
A Cloro request typically takes 30 to 45 seconds, and around 20 seconds for ChatGPT according to public user reviews. That is because Cloro runs a real browser session against the live AI interface for every request rather than calling a model API. The speed is fine for scheduled monitoring and batch jobs, which is what the async endpoints are built for, but rules it out for anything a user is waiting on in real time.
No, unused Cloro credits expire at the end of each monthly cycle. That makes the plans a poor fit for bursty usage, where you run a big audit one month and nothing the next. The upside is that failed requests are never charged and queued async tasks are not billed until they process, so what you do spend goes on results you actually received. The GET /v1/credits endpoint lets you check your balance and reset date programmatically.
Use Cloro when the point is AI answers: ChatGPT, Perplexity, Copilot, Gemini and Google's AI Overview and AI Mode, with citations parsed for you. Use DataForSEO when you need bulk Google SERP data at the lowest cost plus backlinks, keywords and on-page data from the same vendor. Use SerpApi when you need the widest engine list (80-plus, including Bing, YouTube and eBay) and its legal indemnity. DataForSEO and SerpApi do not offer native ChatGPT, Perplexity or Copilot endpoints, and SerpApi bills per batch of 10 results, which gets expensive at depth.
For a raw API, the closest Cloro alternatives are DataForSEO, SerpApi and Serper.dev for Google data, and SearchApi for a mix of Google and AI engines. If what you actually want is a finished AI visibility dashboard, look at Rankscale, Rank Prompt, OmniSEO, HubSpot AEO or the AI visibility features inside Semrush and SE Ranking. Several dashboard products are built on top of APIs like Cloro's, so the choice is really whether you want to build or buy.
Yes. Cloro has integrations for Python, JavaScript, n8n, Zapier and LangChain, plus a hosted MCP server at mcp.cloro.dev that exposes every monitor endpoint as a tool for Claude, Claude Code, Cursor or any MCP client. The MCP server uses your own API key, so billing and concurrency limits are the same as direct requests. For no-code users, n8n or Zapier is the quickest way to turn Cloro output into a Google Sheet or Slack alert.
Cloro is not for marketers who want to log in and read a report, for anyone who needs answers in a few seconds, or for teams whose usage is bursty enough that expiring credits would be wasted. It also inherits every change the AI platforms make: in August 2026, OpenAI's move to a new ChatGPT interface format left several response fields empty, and Perplexity restrictions pushed that endpoint's price up. If that kind of upstream volatility would break your workflow, a dashboard vendor absorbs it for you.
Final Verdict

Should You Try Cloro?

Cloro does one hard thing well: it reads nine search and AI surfaces through a single API and gives you back JSON you can work with straight away, the citations and search queries already separated out. Nothing else I've tested covers the chat engines and Google in one schema at this price, the free tier is enough to build on, and the billing only charges you for results. Public user feedback lines up with my own experience: the relief is in not maintaining the scraper yourself.

The reservations are the ones you'd expect from infrastructure. There's no reporting, so someone has to build it. Requests take half a minute. Credits die at month end. And the platforms it reads keep changing, which means empty fields and breaking changes from time to time, handled quickly but handled nonetheless. For marketers without a developer, a dashboard tool is the better buy. For anyone building AI visibility tracking, whether for a client, a data warehouse or a product, Cloro is the API I'd reach for first in 2026.

Our Recommendation

Start on the free tier. Create a key, run one sync request per endpoint you care about, and look at the JSON before you look at anything else. If the citation pills and sources give you what you need, set up an async batch for your real prompt list and write the results to a sheet or a database. Go to Hobby rather than Lite once you're monitoring daily; the per-credit rate halves and 20 concurrent jobs will keep up with a real prompt list. Subscribe to the changelog RSS, and build your parser to expect empty fields. Do that and you'll know inside a week whether Cloro fits.

Try Cloro →
4.2
★★★★☆
Very Good
About This Cloro Review: Based on hands-on use of a free Cloro account, including live API requests, the dashboard, billing page and playground, plus Cloro's published documentation and changelog, independent write-ups and public user feedback. Published September 2026. This review contains affiliate links, so we may earn a commission at no extra cost to you. Our ratings and editorial opinions remain independent.