Skip to main content
Pay as you go, billed per 1,000 results. You buy credits up front, and each job spends them per result it delivers. Each scraper has its own rate, and every job reports its final cost.

​
Rates

The current rates are on the pricing page, and in the price list, which needs no key:
The price list includes only public scrapers. Each row has platform (the scraper type), price_per_1000 and price_per_item, in USD as decimal strings, plus minimum_charge_enabled and minimum_charge_usd (minimum charge). instagram_posts and instagram_profile bill at the instagram rate; facebook_posts and facebook_profile at the facebook rate. A crawl limited to posts from the last 29 days is billed at its own rate. The dashboard shows that rate before you start the job, and every job keeps the rate it was billed at. The price list shows the rate for a full crawl, and POST /api/v1/pricing/calculate_cost/ estimates at that rate:
A job’s own rate is in price_per_1000_at_creation, and is_initial_crawl says whether it was billed as a full crawl (true) or a recent one (false).

​
What a job costs

A job costs its results times its scraper’s rate, locked when the job is created. A result is one row of output: a post, comment, video, profile, page or search query, depending on the scraper. A new rate applies to new jobs only.
  • You pay per billed result row delivered. A job that delivers zero billed rows costs $0.
  • Some scrapers also return rows that are not posts or videos. Instagram and Facebook start each profile with a summary row, which is billed. Status rows on YouTube and Instagram cost nothing and stay in the results: a channel with nothing in your date range, a link that is not a channel, a private, missing, restricted or blocked profile, a profile that couldn’t be resolved (profile_error), a post that couldn’t be read. Each scraper guide lists its rows and whether they are billed. /results/ total includes these $0 rows and can exceed items_scraped, which counts billed rows.
  • Some scrapers charge a minimum per job. In the price list, minimum_charge_enabled says whether a scraper has one and minimum_charge_usd gives the amount, or null when there is none. A job’s minimum_charge_at_creation says whether its scraper had one when it was created. A job with no billed results costs $0 even then.
  • The charge is settled when a job completes, or, for a job its own spending cap stops, when the stopped run reports the rows it delivered. A job you cancel yourself is not charged.
  • No charge is ever above the job’s maxTotalChargeUsd.
Two rows of the price list, cut to the minimum fields:

​
The final cost

Every job reports what it was charged, in three places that always agree:
  • actual_cost on the job (GET /api/v1/jobs/{id}/), in USD;
  • usageTotalUsd in its webhook, and in the SDK’s run;
  • the job’s line in your billing history and in GET /api/v1/usage/.
estimated_cost on the job is the estimate its credit check used.

​
Credit check: 402

If your balance can’t cover a new job’s estimate, the API refuses it with 402 Payment Required. It also counts the quoted estimated_cost of your organization’s pending and running jobs: those quotes plus the new job’s own must not exceed 10 times your balance. A job stopped by its spending cap still reserves that cap until its final charge is settled. A 402 creates no job, so retrying the request after backing off is safe. See Errors for the response and how to handle it. The estimate is the scraper’s rate times the results the job can return, counting every target: a job for 50 profiles with limit: 100 is quoted as 5,000 results. It is in the job’s estimated_cost. Without a limit, a target is counted as:
  • 50 posts on Instagram and Facebook;
  • 1,000 videos on TikTok, or 100 with a date filter within the last 29 days;
  • 100 videos on YouTube;
  • one row on instagram_profile and facebook_profile.
On Instagram, Facebook, TikTok and YouTube, a job with a date filter within the last 29 days is counted at no more than 50 results per target for each day of its window: a job for 20 channels over the last day, with a limit of 50,000, is quoted as 1,000 results. This lowers only the quote: the scraper still collects the whole window up to your limit, you pay for the rows it delivers, and maxTotalChargeUsd is what caps that. Comment scrapers count maxComments, TikTok search maxItems, TikTok profile its handles and posts, and Google search its queries.

​
Cap a job with maxTotalChargeUsd

Add maxTotalChargeUsd to a job’s input (config): a number of USD, at least 0.01.
maxTotalChargeUsd is a ceiling on spending. It can lower a job’s volume. For tiktok_search without maxItems, the budget sets the volume, up to 100,000 items; other types never run above their volume without a budget.
  1. Split per target. On instagram, tiktok and facebook the cap is split across the profiles or pages and written as a per-target limit (maxPosts, maxVideosPerProfile, resultsLimit), never above the limit you set or the default. A job with no limit gets one from the cap.
  2. A result cap. On instagram_comments and tiktok_comments (maxComments), tiktok_search and tiktok_profile (maxItems), a cap you leave out is set to what the budget buys. tiktok_search is capped at 100,000 items. An explicit maxItems whose cost exceeds the budget is refused with 400; a smaller cap is kept.
  3. When it can’t fit. A job is refused with 400 only when the limit you set for one target already costs more than the whole budget, or the budget buys less than one result per profile. On instagram_profile, facebook_profile and google_serp, the whole job must fit the budget. The message says how many results the budget buys.
  4. YouTube gets no cap written: the job is accepted with a warning in its log, and the budget is enforced while it runs.
  5. While it runs: a job whose results pass 110% of the budget is stopped. Its status becomes cancelled, and its error_message says so. Its actual_cost stays null until the stopped run reports the rows it delivered. It is then billed for them, at most the budget, and its one job.cancelled webhook carries that charge. A run that hasn’t reported 30 minutes after the stop is closed with no charge. A job its budget stopped can’t be retried (400), billed or not: create a new job instead.

​
Add credits

Write to support@go-fetch.io and we’ll add them to your organization. Your balance is credit_balance in GET /api/v1/organizations/.

​
Low balance

When your balance falls below its warning threshold, or to zero, the organization’s owners and admins get an email, and webhooks subscribed to account.low_balance are called. Set the threshold in the dashboard’s organization settings.

​
Usage

GET /api/v1/usage/ lists one usage record per completed job: job, platform, posts_scraped (results billed), unit_price and total_cost, newest first. GET /api/v1/organizations/{id}/usage/ sums the current month.