Pricing

from $1.00 / 1,000 operations

Try for free

Go to Apify Store

AI Token Saver — Deterministic Tools for AI Agents

Try for free

Cut LLM token spend: 80+ deterministic text, web and data operations (HTML to Markdown, chunking, PII redaction, JSON repair, extraction, dedupe, math, dates) for AI agents, with per-call token-savings estimates.

Pricing

from $1.00 / 1,000 operations

Rating

0.0

(0)

Developer

Justin Hagerty

Actor stats

Bookmarked

Total users

Monthly active users

13 days ago

Last modified

Categories

Developer tools

Operation

operation

Optional

Operation to run, e.g. html_to_markdown, chunk_text, redact_pii, repair_json, table_query, math_eval, cron_next_runs. Call list_operations (free) for one-line summaries or explain_operation for options. Leave empty when using pipeline.

Type:string

Options:

extract_headingsextract_imagesextract_linksextract_main_contentextract_metadataextract_tablesfetch_urlhtml_to_markdownhtml_to_textchange_casechunk_textclassify_by_rulesclean_textcompress_contextcount_tokensdeduplicatedetect_languagediff_textextractive_summarykeywordsnear_duplicatesnormalize_whitespacereadabilityregex_extractsentimentslugifysplit_sentencesstrip_markdowntext_statstruncate_to_tokensextract_code_blocksextract_datesextract_emailsextract_entitiesextract_jsonextract_key_valuesextract_moneyextract_numbersextract_phone_numbersextract_urlsdetect_secretsredact_piirestore_piiscan_prompt_injectioncoerce_to_schemaconvert_formatdedupe_recordsdiff_jsonflatten_jsonfuzzy_matchinfer_json_schemajson_queryjson_to_pydanticjson_to_typescriptmerge_jsonnormalize_recordspivotrepair_jsonsample_recordsstatisticstable_queryunflatten_jsonvalidate_json_schemacompound_interestconvert_unitsloan_paymentmath_evalpercent_changeround_numbersconvert_timezonecron_explaincron_next_runsdate_adddate_diffduration_parse_formatformat_datetimeparse_datetimeunix_timesemver_comparevalidateencode_decodeexplain_operationgenerate_uuidhash_textlist_operationsrandom_sample

Input text

input

Optional

Text, HTML, Markdown or JSON text the operation works on (single mode; also $input of a pipeline). Max 5 MB.

Type:string

URL

url

Optional

Public http(s) URL for fetch_url (or the first step of a pipeline). Fetched with SSRF protection; robots.txt respected by default.

Type:string

Data (JSON)

data

Optional

JSON input for data operations (validate_json_schema, table_query, convert_format...). To pass an array, wrap it: {"items": [...]} — or put JSON text in Input text.

Type:object

Options

options

Optional

Operation options as JSON, e.g. {"maxTokens": 500} for chunk_text. Unknown options are rejected with a hint; see explain_operation.

Type:object

Batch inputs

inputs

Optional

Batch mode: run operation with the same options on each item (strings, or JSON values for data operations). Each successful item is charged. Max 10,000.

Type:array

Pipeline

pipeline

Optional

Pipeline mode: steps [{"operation": ..., "options": {...}, "input": "$prev"}]. Each step gets the previous result by default; references: $input, $prev, $steps[0].field. One charge per executed step. Leave Operation empty.

Type:array

Pipeline output

pipelineOutput

Optional

last: one item with the final step's result (default, leanest). all: one item per step.

Type:string

Default:last

Options:

lastall

Include token-savings stats

includeStats

Optional

Add stats (input/output token estimates, LLM-equivalent tokens, estimated tokens and USD saved, ms) to each item.

Type:boolean

Default:true

Compact output

compactOutput

Optional

Drop null fields from results to save your agent tokens.

Type:boolean

Default:true

LLM pricing for savings estimates

llmPricing

Optional

USD per million tokens of the model your agent would otherwise use: {"inputPerMTok": 3, "outputPerMTok": 15}. Only affects usd_saved_est.

Type:object

Default:

{
  "inputPerMTok": 3,
  "outputPerMTok": 15
}

Respect robots.txt

respectRobotsTxt

Optional

fetch_url obeys robots.txt (RFC 9309). Disable only for sites you are entitled to fetch.

Type:boolean

Default:true

Fail run at severity

failOnSeverity

Optional

CI gate shared by our Actors. This Actor emits no scored findings, so it has no effect here.

Type:string

Default:none

Options:

nonecriticalhighmediumlow

Disabled rules

disabledRules

Optional

ATS rule IDs whose warnings to suppress (see RULES.md). Requires the custom_rules feature (Pro plan and above).

Type:array

License key

licenseKey

Optional

Optional. Enterprise / prepaid license key. Store users don't need one.

Type:string

Product Data Extractor API — Price, Title, Images & Stock

highbrow_fame/product-page-extractor-api

Name, brand, price, currency, stock, rating, images, SKU/GTIN and variants from a product page's schema.org JSON-LD, microdata or Open Graph tags. One JSON record per URL. Product Data Extractor API.

yestrue

PDF to Markdown API — Text, Headings & Metadata from Any PDF

highbrow_fame/pdf-to-markdown-api

Converts PDFs at public links to Markdown and plain text with page markers, headings, links and metadata; scanned pages are flagged, never charged. Unofficial PDF to Markdown API.

yestrue

Website to Markdown Crawler for LLM, RAG & AI Context

physealabs/website-markdown-crawler

Crawl any website and get each page as clean Markdown with title, description, word count and token estimate. Menus, footers and cookie banners removed. Built by Physea Labs (Liminal). Launch Sale: 90% OFF for the first week. $0.001 per page.

jay casey

PDF & Image to Text / Markdown (with OCR)

singgendarhan/pdf-image-to-text-markdown-ocr

Convert PDFs and images to clean Markdown and text for RAG, AI agents and automations. Keeps headings, lists, tables and page breaks; OCRs scanned pages and images automatically (28 languages). Pay only for pages that produce text.

Nick Na

SiteVerdict — robots.txt & Sitemap Monitor

siteverdict/siteverdict-robots-sitemap-monitor

Catch SEO disasters before Google does: watches robots.txt, XML sitemaps and key pages for Disallow: /, noindex, canonical and status changes, sitemap URL drops and errors. Diffs vs last run, Slack/webhook alerts, HTML report.

Aman Thakran

TikTok Transcript Scraper: Video to Text with Timestamps

scrape-lads/tiktok-transcript-scraper

TikTok Transcript Scraper: get the transcript of any public TikTok video or a creator's latest videos. TikTok's own captions first, Whisper speech-to-text when there are none. Timestamped segments, language, caption and stats.

Scrape Lads

AI Web Data Extractor – Any URL or PDF to JSON, No API Key

gazidev/ai-web-data-extractor

Turn any web page, PDF or DOCX into structured JSON with an LLM. Describe the fields in plain English or a JSON schema; output is validated. No OpenAI/OpenRouter key: tokens are billed to your Apify account via Apify's OpenRouter Actor. $4 per 1,000 pages.

Cemal Atakli

PDF to Markdown, Tables & RAG Chunks (with OCR)

errandkit/pdf-to-markdown-tables

Convert PDF, DOCX and image files to clean Markdown, tables as CSV and RAG-ready chunks with page numbers. Scanned pages are read with OCR. No API keys, pay per page, built for AI agents and LLM pipelines.

Killis Haraway

PDF to Markdown & Text Extractor for LLM and RAG

datagleaner/pdf-to-markdown

Convert PDF URLs to clean Markdown or plain text for LLMs, AI agents and RAG. Headings, lists, paragraphs and simple tables, optional ready-to-embed chunks with page numbers. No start fee, $0.002 per PDF, scanned PDFs are not charged.

Data Gleaner

MCP Tools Token Profiler and Schema Linter

leadharbor/mcp-tools-token-profiler

Paste an MCP tools/list or give public server URLs. Get estimated tokens per tool, the 5 heaviest tools, ranked savings with exact JSON paths, and lint cited to the MCP spec. No login.