Pricing
Free AI visibility APIs in 2026: start with Bright Data and cloro
Eleven of the 55 AI visibility APIs we rate hand out a key with no payment, from 125 searches to 5,000 AI answers a month. None returns share of voice.

By Mathias C.SEO consultant, and more recently GEO, at MentionBox
Eleven of the 55 AI visibility APIs this site rates will hand you a working key with no payment, and the allowances are not close to each other: Bright Data renews 5,000 credits a month, which is 5,000 scraped AI answers at one credit per record, while Scale SERP renews 125 Google searches. Between those two sit cloro's 500 credits, Nimble's 5,000 requests, SERPHouse's 4,000 credits, HasData's 1,000 credits, Apify's $5 of platform usage and SerpApi's 250 searches. None of the eleven returns share of voice, a brand position or sentiment. Every free tier in this market is a collection tier, and the counting is yours to write. API for GEO is our editorial pick after hands-on testing by Mathias C., and it sits on the other side of that line: no free credit amount is published, and the metrics arrive computed at $13 per 1,000 answers. Prices checked on October 23, 2026.
This guide is for a small business or a one-developer team that wants AI answer data this week without raising a purchase order, and for anyone who typed "free AI visibility API" and landed on a page listing free dashboards instead. It separates the three offers that all get called free, prices each monthly allowance in answers rather than in credits, works through a real 20-prompt list, reads the resale clause that decides whether free data can reach a client, covers the platforms whose free plan stops at the dashboard door, and ends with the one free route through an official engine API.
Free tier, free credit, free trial: three different offers
Vendors use one word for three things, and the difference decides whether you can still collect data in March.
| Offer | What renews | Example in this market | What it is for |
|---|---|---|---|
| Standing free tier | an allowance, every month | Bright Data, 5,000 credits a month | a small list tracked for as long as you like |
| One-off free credit | nothing, it is spent once | DataForSEO, $1 of trial credit | one afternoon of reading the response schema |
| Timed trial | nothing, it ends on a date | Profound, 50 prompts daily for 7 days | a demo of a dashboard, often without the API |
Offers read on each vendor's pricing page, checked on October 23, 2026.
Three consequences follow, and they are why this page splits the eleven from the rest:
- A trial cannot run a tracker. Seven or fourteen days of collection gives one reading per prompt, and engines resample their own answers, so a series that short shows variance rather than movement. Our explainer on why the same prompt gives different answers sets out how wide the spread gets.
- A one-off credit is a schema reader. DataForSEO's $1 buys about 833 LLM Scraper results at $0.0012 each, enough to learn the payload and not enough to chart anything, and the next step is its $50 minimum payment.
- A free tier is a budget line of zero. Twenty prompts on three engines, run weekly, is 260 answers a month. Several of the eleven hold that indefinitely.
The eleven free tiers, side by side
The table prices each free allowance in AI answers, using the credit cost per engine that each vendor publishes rather than the credit count on the pricing card.
| API | Free offer, per month | Free AI answers a month | Engines on the free key | Card needed | What the response carries | Resale |
|---|---|---|---|---|---|---|
| API for GEO | sandbox project with the key, no free credit published | not published | 10 | no, a form | answer, ranked sources, position, share of voice, sentiment, themes, fan-out, shopping | allowed |
| cloro | 500 credits, 1 concurrent job | 100 ChatGPT, or 125 Gemini | 6 | no | answer, cited URLs, entities, fan-out queries, shopping | forbidden |
| Bright Data | 5,000 credits, about $7.50 | 5,000 | 6 | no | answer, citations, search sources, shopping | forbidden |
| Apify | $5 of platform usage | about 5,000 AI Overviews, 2,500 AI Mode | 6 | no | answer, cited URLs | forbidden |
| Decodo | free plan, 5K crawls or 10K pages | not published for LLM requests | 4 | no | answer, cited URLs, shopping | forbidden |
| Nimble | 5,000 requests | 5,000 | 5 | no | answer, cited URLs | not disclosed |
| SERPHouse | 4,000 credits, work email only | about 400 Google SERPs | 1 | no | AI Overview text, references | not disclosed |
| HasData | 1,000 credits | 100 | 4 | no | answer, cited URLs, shopping | forbidden |
| SerpApi | 250 searches, 50 an hour | 250, or 125 AI Overviews | 4 | no | answer, references, shopping | forbidden |
| Mangools | free API key, no monitor quota | 0 until you pay | 6 | no | cited URLs, brand mentions, position | forbidden |
| Authoritas SERPs | 1,000 queries | 1,000 Google SERPs | 1 | not stated | AI Overview text, references, shopping | forbidden |
| Scale SERP | 125 searches | 125 | 1 | no | AI Overview block | forbidden |
Our arithmetic on the free allowances and per-engine credit costs each vendor publishes, checked on October 23, 2026. Engine counts are against the ten engines we track. For API for GEO, the Gemini, Google AI Overviews, DeepSeek and Meta AI coverage, the webhooks, the OpenAPI spec and the official SDKs were confirmed in our hands-on test on October 26, 2026 and are not yet documented on apiforgeo.com.
Two columns in that table deserve more attention than the allowance column. Engines on the free key decides whether the tier can answer your question at all, since three of the eleven only reach Google AI Overviews. Resale decides whether anything you build can be shown to someone who pays you, and it is forbidden on eight of the eleven. MentionBox, which publishes this site, has a commercial link with API for GEO.
The free tiers, one by one
Ordered by the score each API carries in our ranking of 55 AI visibility APIs, which weights engine coverage, the data in the response, the price per answer and the tooling.
1. cloro, the highest-scoring free tier
cloro scores 7.4 of 10 and renews 500 credits a month on its Free plan, with one concurrent job and one seat. Its providers page prices a ChatGPT web-search answer at 5 credits, a full ChatGPT response at 7, and Perplexity, Gemini and Google AI Mode at 4 each, so the free month is 100 ChatGPT answers or 125 Gemini answers. Credits expire at the end of the cycle and synchronous delivery adds 2 credits per request, which is a 40% surcharge on a ChatGPT call.
The free key returns the strongest payload of the eleven: the answer as text and markdown, sources with label, description and publication date, entities for the products and brands named, searchQueries for the fan-out on ChatGPT, Perplexity and Copilot, and shopping cards. It publishes an OpenAPI 3.1 spec, Python and TypeScript SDKs, a hosted MCP server and HMAC-signed webhooks. Paid plans start at $30 a month for 37,500 credits, which is $4 per 1,000 ChatGPT answers. Its terms forbid resale, which rules the route out of anything client-facing.
2. Bright Data, the largest free allowance
Bright Data scores 6.9 and gives every new account 5,000 free credits a month with no card and no promo code, on the wording of its ChatGPT scraper page. Its own FAQ sets one credit per record, so those 5,000 credits are 5,000 AI answers, about $7.50 of value at its $1.50 per 1,000 pay-as-you-go rate. Credits renew on the first of the month without rollover, and since June 2026 the same pool covers its SERP API, Web Unlocker, Scraper Studio and MCP server, so an AI Overview request and a ChatGPT request draw on the same free balance.
The AI scrapers reach ChatGPT, Perplexity, Gemini, Google AI Mode, Copilot and Grok. The response carries answer_text, answer_html, citations with title and description, and search_sources with a rank, plus shopping cards on ChatGPT and Perplexity. There is a Python SDK, a /trigger endpoint taking up to 5,000 URLs per call, webhooks and delivery to S3, GCS, Pub/Sub, Snowflake and SFTP. Past the free credits, requests are refused until funds are deposited, and its licence forbids reselling the service without written authorisation. Our duel cloro vs Bright Data compares the two payloads field by field.
3. Apify, five dollars of usage a month
Apify scores 6.6 and its Free plan carries $5 of platform usage a month. The dedicated scrapers price per event: its Google AI Overviews Scraper publishes from $1.00 per 1,000 overviews, the AI Mode scraper from $2.00 and the Perplexity scraper from $9.00, so the free month is roughly 5,000 AI Overviews, 2,500 AI Mode answers or 555 Perplexity searches. One caution: Apify's own blog states that the AI features of the general Google Search Results Scraper need a paid plan, so the free month runs on the dedicated actors.
The actors return the answer and the cited URLs, with JavaScript and Python clients, an OpenAPI tab, batch runs and an MCP server, and no brand figures at all. Paid plans start at $19 a month, and the terms forbid sublicensing or redistributing the service. An actor run is a job to schedule rather than a request to make, so the free tier suits a one-off study of Google's AI surfaces better than a tracker.
4. Decodo, a free plan with an unpublished answer count
Decodo scores 6.4 and offers a free plan on its web scraping page of up to 5,000 crawls or 10,000 pages scraped with no card. It does not publish how many LLM template requests those crawls represent, and its ChatGPT page advertises a separate seven-day free trial.
Its LLM templates cover ChatGPT, Perplexity, Gemini and Google AI Mode, returning the answer, the cited URLs and shopping blocks, with a TypeScript SDK and an MCP server, and resale is forbidden. Treat the free plan as a trial of the mechanics and price the real work at the $1 per 1,000 of its $19 monthly plan, among the cheapest in our ranking of the cheapest AI visibility APIs.
5. Nimble, 5,000 free requests and a steep first plan
Nimble scores 5.7 and renews 5,000 free requests a month under a "try every API for free" offer with no card, on its pricing page. One template execution is one request, so the free month is 5,000 AI answers, the joint-largest allowance in this set, worth $15 at its $3 per 1,000 pay-as-you-go rate. Its AI templates reach ChatGPT, Perplexity, Gemini, Google AI Overviews and Grok, and return the answer and the cited URLs with Python, TypeScript and Go SDKs.
The cliff is what to plan for. Nimble's first subscription is Startup at $2,500 a month billed annually, so a team that outgrows the free requests moves either to pay-as-you-go at $3 per 1,000 or to a five-figure annual commitment. Resale terms for the API are not disclosed on the pages we read, which for client work amounts to the same answer as a no.
6. SERPHouse, 4,000 credits if your email has a domain
SERPHouse scores 5.0 and is the one free tier in this set whose renewal depends on your email address. Its pricing page gives the Free plan 4,000 API credits, renewing monthly on a work email and lasting seven days without auto-renewal on a personal one. It does not publish a credits-per-request figure for the free plan; its Basic plan at $29.99 carries 400,000 credits with SERP requests at $0.75 per 1,000, which works out at about 10 credits a request, so 4,000 free credits is roughly 400 Google SERPs a month.
Coverage is Google AI Overviews only, returned as ai_overview.contents with ai_overview.references carrying title, link, description and source, through Node.js, PHP and Python clients, a hosted MCP server with 23 tools and a scheduled async mode. Its terms carry no general resale clause, but section 21 forbids reselling, redistributing or sublicensing data obtained during the trial, which covers the free period described above.
7. HasData, a hundred answers a month, every month
HasData scores 4.5 and renews 1,000 free credits a month with no card, with every tool included. Its AI Mode quickstart sets 10 credits per request across AI Mode, AI Overview, ChatGPT and Perplexity, so the free month is 100 AI answers, and credits are only charged on success. An AI Overview can cost a second 10-credit call when a page token is needed, which halves that figure on Google's late-loading overviews.
The response gives textBlocks typed as paragraph, heading, list, table, code, quote, shopping results and local results, plus references with link, title, snippet and index, through an MCP server with four AI tools and no published SDK. Paid plans start at $49 a month for 200,000 credits, which is $2.45 per 1,000 answers, and its terms forbid reselling the data or reselling access to it through your own paid plan without written permission. A hundred answers a month is a genuine monitor for a single business with five prompts on two engines, checked weekly.
8. SerpApi, 250 searches a month with ten client libraries
SerpApi scores 4.3 on price and allowance, and higher than anything else here on tooling. Its pricing page gives the Free plan 250 searches a month at 50 an hour, and counts only successful searches, with cached and failed searches excluded. One AI Mode, Gemini or Copilot answer is one search; a Google AI Overview loaded late through a page token is two, so the free month holds 125 of those.
Coverage is Gemini, Google AI Overviews, Google AI Mode and Copilot, parsed into text_blocks, a reconstructed_markdown field and references listed in the order the answer cites them, with shopping results on AI Mode and Gemini, client libraries in ten languages and an MCP server. The weakness is the rate past the free plan: $25 a month for 1,000 searches is $25 per 1,000 answers, the highest among the eleven, and $50 per 1,000 on late-loading AI Overviews. The detail of the Google surfaces sits in SerpApi vs DataForSEO for Google AI Overviews data.
9. Mangools, a free key with no data behind it
Mangools scores 3.6 and is the odd one in this set. Its API page offers an API key free with any account and marks the AI Search Watcher endpoints free with no lookup quota, while creating the monitors that generate the data consumes paid quota. The free key reads data you have to pay to produce, which is the inverse of every other tier here.
It is worth knowing for what those endpoints return: cited URLs, brand mentions and a position, across ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Claude and Grok, with GET /aiwatcher/models reporting per model whether it is active, deprecated, supports citations and is a SERP surface. That is the only free-key route in our set that hands back a brand position rather than raw text. It publishes no price per answer, and resale is forbidden.
10. Authoritas SERPs API, a thousand Google queries a month
Authoritas scores 3.3 and its API pricing page opens with 1,000 free queries a month to test the API. One SERP is one query and only successful responses are billed, so the free month is 1,000 Google result pages with the AI Overview requested through its include_sge parameter. Coverage stops at Google AI Overviews, and the response carries the overview text, its references and shopping blocks, through a PHP client and no MCP server.
Two limits to weigh. The next tier up is £300 a month for 100,000 queries at £0.003 a query, and the page states that packages are not yet orderable online, so growing out of the free tier means a sales conversation. Its acceptable use policy states that no other person may use the API services by assignment, licence or permission.
11. Scale SERP, 125 searches and a Google-only payload
Scale SERP scores 2.4, the lowest in this set, and renews 125 free searches a month with no card on its pricing page. It returns the Google AI Overview block, computes nothing, and its paid plans start at $66 a month for 10,000 credits, which is $13.20 per 1,000 AI Overviews, nearly nine times Bright Data's rate. Resale is forbidden. Four answers a day is a connectivity test rather than a series.
Does a free tier hold a real prompt list?
Take the smallest list that teaches a small business anything: 20 prompts, three engines (ChatGPT, Gemini and Google AI Mode), one country, run once a week. That is 60 answers a cycle and 260 a month, counting 52 weeks across 12 months.
| Free tier | Covers all three engines | Free answers a month | Holds the 260-answer list |
|---|---|---|---|
| Bright Data, 5,000 credits | yes | 5,000 | yes, with 4,740 to spare |
| cloro, 500 credits | yes | 100 ChatGPT or 125 Gemini | no, the list costs 1,127 credits a month |
| Nimble, 5,000 requests | no, Google AI Mode missing | 5,000 | two of three engines |
| SerpApi, 250 searches | no, ChatGPT missing | 250 | two of three engines, 173 answers |
| HasData, 1,000 credits | no, Gemini missing | 100 | no, about a third of the list |
| Apify, $5 of usage | no, ChatGPT needs a paid plan | about 2,500 AI Mode | yes on Google's surfaces |
Our arithmetic on the published allowances and per-engine credit costs, checked on October 23, 2026. The cloro figure is 20 prompts × (5 + 4 + 4) credits × 52 ÷ 12 weeks.
A small list genuinely fits in a free tier, which is the honest answer to the query that brought you here, and the engine column decides it more often than the allowance does: Bright Data's 5,000 credits cover the list nineteen times over, while cloro's richer payload runs out in the second week. The figure teams meet in month two is the other one, when the cadence goes from weekly to daily and 260 answers becomes 1,800.
What no free tier gives you
Every one of the eleven hands back the answer and the sources, and nothing else. Four figures stay on your side of the line, and each one is a decision rather than a calculation:
- Brand matching. Aliases, legal suffixes, product names, possessives and misspellings decide whether a mention counts, and a plain string search produces a share of voice that is plausible and wrong.
- Position. The rank of a mention inside an answer needs one rule for lists, one for prose and one for tables, and the rule has to stay fixed or the series breaks.
- Share of voice. Your mentions over all brand mentions requires a competitor set you maintain by hand, because a new entrant appearing in 40% of answers will not add itself.
- Sentiment. A model call per answer, with its own cost and its own drift the day the model behind it is updated.
Keep the raw answers while you are at it, so a change in the counting rule costs a recount rather than a fresh month of collection. GEO metrics explained sets out the choices you make when you write that code, and how to build an in-house GEO dashboard prices the build.
Can free-tier data reach a client?
This is the question that disqualifies more free tiers than any allowance, and it is answered in the terms rather than on the pricing page.
- Forbidden outright: Bright Data requires prior written authorisation to resell the service, cloro forbids selling, reselling, licensing, sublicensing or distributing the services, HasData forbids reselling the data or access to it through your own paid plan, SerpApi forbids exploiting any portion of the service without written permission, and Authoritas states no other person may use the API services.
- Forbidden for the free period specifically: SERPHouse bars reselling, redistributing or sublicensing data obtained during the trial, which is the free plan.
- Not disclosed: Nimble publishes site terms rather than API resale terms, which leaves an agency carrying the risk.
- Allowed: API for GEO, whose terms state that reselling is an intended use and whose pricing page lists data resale and white label as yes, while forbidding sublicensing direct API access.
An internal dashboard is fine on any of these tiers. A figure in a client deck, a white-label report or a GEO feature inside your own product is not, and no upgrade changes that for the eight that forbid it. Our guide to white-label AI visibility APIs lists the routes that permit it.
Why the AI visibility platforms have no free API
Search for a free AI visibility API and most results are dashboards with a free plan. The pattern across the platforms we rate is consistent: the free plan shows you numbers on a screen, and the key that would let you fetch them sits one or two tiers up.
| Platform | Free offer | API on that offer | What the API costs |
|---|---|---|---|
| AthenaHQ | free Essential plan with $25 of credit | no | paid add-on on Starter at $295 a month, or Enterprise |
| Profound | 50 prompts daily for 7 days, 3 engines | no, listed as None | Enterprise plan, price on quote |
| ZipTie | 7-day trial, 25 daily prompts, no card | no | $10 a month add-on, on a paid plan only |
| Rankscale | 7-day Pro trial | no | Pro+ at $229 a month |
| LLM Pulse | 14-day trial on the weekly plans | no | Scale at €299 a month, MCP on every plan |
| Scrunch | 7-day trial of Starter or Core, 100 prompts | no | Agency Core at $500 a month, or Enterprise |
| Peec AI | none published | no | Enterprise, billed annually |
| Promptwatch | 7-day trial | yes, on paid plans | included from $95 a month |
| Rankability | 7-day trial, card required | yes, on paid plans | included from $97 a month |
Plan and documentation pages read on October 23, 2026. AthenaHQ's documentation states that API access is a paid feature and that the page is restricted entirely on the Free plan.
Two of the nine include the API on every paid plan, and they are worth naming for it: Promptwatch at $95 a month on Essential and Rankability at $97 on Starter both ship API and MCP access on the entry tier. Keyword.com goes further down the price scale, with a 14-day card-free trial and an AI Visibility plan from $9.80 a month, though that plan carries 20 credits. For the rest of the group, a dashboard trial says nothing about key access, and our guide to Profound alternatives with a self-serve API goes through the seven routes that publish both a key and a rate.
The free route through an official engine API
One engine gives away grounded answers at a volume that counts. Grounding with Google Search in the Gemini API carries 5,000 free search requests a month, shared across Gemini 3 and newer models, then $14 per 1,000 requests, under a default limit of 1,500 requests a day per project. The response returns the google_search_call steps showing which queries the model ran and inline url_citation annotations with the source title and the character positions they support, which is the fan-out and the citation list in one payload.
| Official API | Free allowance | Price after it |
|---|---|---|
| Gemini, Grounding with Google Search | 5,000 search requests a month | $14 per 1,000 requests |
| OpenAI web search tool | none | $10.00 per 1,000 calls, plus search content tokens |
| Anthropic web search tool | none | $10 per 1,000 searches, plus token costs |
| Perplexity Search API | none published | $5.00 per 1,000 requests, $1.00 for Fast Search |
| Google AI Mode, Google AI Overviews | no official API | not available at any price |
Rates read on each provider's pricing page on November 19, 2026.
The catch applies to every official-API route and not only to Google's. An API answer is not the answer the consumer product gives: the retrieval, the system prompt, the personalisation and the model version differ, and the model only searches when it decides to. Our explainers on Gemini API grounding and on the OpenAI web search tool against ChatGPT answers go through what changes. For a small brand watching Gemini, 5,000 free grounded requests a month is still the cheapest honest signal available, as long as the report says which surface it came from.
Five mistakes on a free AI visibility tier
- Reading the credit count as an answer count. A 1,000-credit allowance is 100 answers at HasData and 1,000 at Bright Data. The multiplier is on the provider's own docs page, never on the pricing card.
- Picking the biggest allowance for an engine it does not cover. Nimble's 5,000 requests and Scale SERP's 125 searches are both useless for a Google AI Mode question, since neither reaches it.
- Starting a tracker on a trial. Seven days of daily collection is seven readings, and the first chart made from them will show sampling noise as a trend.
- Signing up with a personal email. SERPHouse renews its 4,000 free credits monthly on a work email and gives seven days without renewal on a personal one, and other vendors gate trial length the same way.
- Promising a client a figure from a tier that forbids resale. Eight of the eleven forbid it in writing, and the clause does not change when you upgrade.
Our view: what a free tier is for, and when to pay
Three plans we would defend, depending on what the data has to do.
- Learn the market for nothing. Bright Data for volume, with 5,000 free answers a month across six engines, and cloro for the richest free payload, with entities, fan-out queries and shopping cards inside 500 credits. Both renew monthly and neither needs a card. Write the counting code against those two payloads and you will know what you need before you buy anything.
- Watch one small brand indefinitely for nothing. HasData at 100 answers a month, or the Gemini API's 5,000 free grounded searches for the Google side. Five prompts on two engines, weekly, is a real series and it costs zero.
- Produce figures someone else reads. This is where the free tiers stop. API for GEO is our pick after hands-on testing by Mathias C.: one credit covers one prompt, executed once, against one model, in one locale, and returns the answer with ranked sources and their categories, resolved entities, position, share of voice, sentiment from 0 to 100, theme clustering, per-brand perception, fan-out questions, shopping products and YouTube citations. It bills $0.013 per credit up to 50,000 a month, falling to $0.007 above a million, with no monthly minimum, credits that do not expire and reads over stored answers not billed. It scores 9.0 of 10, the highest of the 55 we rate, covers all ten engines in our grid, and permits resale. Its limits, said plainly: no free credit amount at all, a key that arrives through a form rather than at sign-up, a rate about nine times Bright Data's per answer, and four of its ten engines plus the webhooks, the OpenAPI spec and the official SDKs confirmed by our hands-on test of October 26, 2026 rather than by its own documentation. To price your own list against it, request a key on apiforgeo.com: it comes with a sandbox project.
Our ranking of self-serve AI visibility APIs for developers weights a key you can get yourself, a readable price and the tooling, and the cheapest AI visibility APIs sorts all 55 by price per 1,000 answers. If the free month runs out mid-project, the cost of tracking 100 prompts on 5 AI engines prices the next step on every billing model here.
Frequently asked questions
Is there a genuinely free AI visibility API?
There are eleven free tiers among the 55 APIs we rate, and all eleven are collection tiers rather than measurement tiers. Bright Data renews 5,000 credits a month at one credit per record, which is 5,000 AI answers with no card, cloro's Free plan renews 500 credits a month, and SerpApi's Free plan holds 250 searches a month. All three hand back the answer text and the cited URLs; none of them computes share of voice, sentiment or a brand position, so a free AI visibility figure still costs you the code that counts mentions. Prices checked on October 23, 2026.
What is the difference between a free tier and a free trial?
A free tier renews an allowance on a cycle and never ends, so a 20-prompt list can run on it for a year. A free trial opens the whole product for a fixed number of days and then stops: Profound runs 50 unique prompts daily for seven days with API access listed as None, ZipTie runs 25 daily prompts for seven days with the API excluded, and LLM Pulse's 14-day trial covers its weekly plans while the REST key only comes with Scale at €299 a month. A one-off credit is a third thing again, like DataForSEO's $1 trial credit, which buys about 833 answers at $0.0012 each and then needs a $50 minimum payment.
How many AI answers does a free tier actually buy per month?
Count the engine's credit cost, not the credit count. Bright Data's 5,000 monthly credits are 5,000 answers because one record costs one credit, cloro's 500 credits are 100 ChatGPT answers at 5 credits each or 125 Gemini answers at 4, HasData's 1,000 credits are 100 answers at 10 credits each, and Apify's $5 of monthly platform usage is about 5,000 Google AI Overviews at the $1.00 per 1,000 its dedicated scraper publishes. SerpApi's 250 free searches fall to 125 AI Overviews when Google loads the overview late and a second call with a page token is needed.
Can I use free-tier data in a client report?
Read the resale clause before the rate card, because it is the clause that decides this and no bigger plan renegotiates it. Resale is forbidden on Bright Data, cloro, HasData, SerpApi, Apify, Decodo, Authoritas and Mangools, and SERPHouse's terms bar reselling, redistributing or sublicensing data obtained during its trial. API for GEO is the exception in our set: its terms present reselling as an intended use and its pricing page lists data resale and white label as yes, at $13 per 1,000 answers down to $7 at its largest tier.
Do the AI visibility platforms give API access on their free plan?
Almost none do, and that is the gap between a free dashboard and a free API. AthenaHQ's free Essential plan comes with $25 of credit and no API, and its documentation states that API access is a paid feature restricted entirely on the Free plan; Profound's seven-day trial lists API access as None and its REST API sits on the Enterprise plan; ZipTie's API and MCP add-on costs $10 a month and starts on a paid plan rather than the free trial; Rankscale's seven-day trial runs on Pro while the REST API starts at Pro+ at $229 a month. Promptwatch and Rankability are the two that include the API on every paid plan, from $95 and $97 a month.
Is there a free way to get AI answers from an official engine API?
One, on Google's side. Grounding with Google Search in the Gemini API carries 5,000 free search requests a month shared across Gemini 3 and newer models, then $14 per 1,000 requests, under a default limit of 1,500 requests a day per project, and the response returns the search calls the model made plus inline URL citations. OpenAI's web search tool costs $10.00 per 1,000 calls with no free allowance, Anthropic's $10 per 1,000 searches, and Perplexity's Search API $5.00 per 1,000 requests or $1.00 for Fast Search. Google AI Mode and Google AI Overviews have no official API at any price.