158 Services from api.relaystation.ai

Pulls specified fields out of any input text and returns them as clean, schema-shaped JSON using a cost-efficient LLM.

Captures a screenshot of any public web page as PNG, JPEG, or WebP using headless Chromium with SSRF protection

Performs full KYC in a single API call — verifying a government-issued ID, running liveness detection, and matching the applicant's face to the ID photo.

Converts raw HTML into clean plain text by stripping all markup, suitable for LLM input preprocessing or search indexing.

Run SELECT SQL queries over large uploaded datasets (CSV/JSON), paying only per byte scanned at $0.01 per 100 MB

Analyzes an image URL and flags unsafe or inappropriate content using AWS Rekognition, returning moderation categories with confidence scores

Extends an image's canvas on any side (top, bottom, left, right) by a specified number of pixels, filling the new area with a background color, copy, repeat, or mirror of the original.

Converts documents between markdown, HTML, docx, odt, rtf, epub, LaTeX, and rst formats using pandoc, with citation support

Overlays or watermarks one image onto another with configurable position and blend control

Permanently burns styled subtitle/caption tracks into a video file using libass, so captions are hardcoded into every copy of the video.

Rasterizes individual PDF pages to PNG or JPEG images, suitable for previews, thumbnails, and vision-model input

Creates a named, versioned, cryptographically-provable storage object (baton) that agents can write to, read from, append, and hand off to other agents or people

Performs a three-way merge of two edited versions of a text against a common base, inserting conflict markers where edits overlap.

Sanitizes HTML against a safe whitelist by stripping scripts, event handlers, and XSS vectors before embedding user-generated content

Recognizes celebrities in an image and returns their names with confidence scores using AWS Rekognition

Extracts form fields as structured key→value pairs from scanned or digital documents using AWS Textract Forms, returning clean JSON without requiring a template.

OCR a document and return full line- and word-level layout coordinates (via AWS Textract) for position-aware text extraction and downstream parsing.

Detects personal protective equipment (face covers, hand covers, head covers) worn by people in an image, returning labeled findings with confidence scores.

Generates a QR code with a custom logo composited in the center using high error-correction (ECC H) to ensure reliable scanning

Calculates a route between two geographic coordinates, returning distance, duration, and geometry as JSON

Converts legacy, exotic, or damaged Office files (doc, xls, ppt, odt, etc.) to PDF using LibreOffice, handling formats that modern Office applications refuse to open

Strips EXIF and metadata from images for privacy while auto-orienting photos to preserve upright orientation

Send a PDF, Word doc, or Markdown file for legally binding e-signature with hosted signing links, ordered multi-signer routing, approvers, templates, and live status tracking.

Generate QR codes as PNG or SVG images with configurable size, error-correction level, and margin

Adds an invisible OCR text layer to a scanned PDF, making it fully text-searchable without altering its visual appearance

Programmatically fill or flatten PDF forms by submitting field values and receiving a completed PDF document in return.

Stamps a text watermark (e.g. DRAFT, CONFIDENTIAL, or custom text) across pages of a PDF document with configurable position, opacity, color, and size

Translates text to a target language preserving meaning and tone, at high accuracy (~Opus-4.8 level) for roughly 1–5% of the cost

Condenses any input text into its key points — briefs, digests, or TL;DRs — at ~Opus-4.8 accuracy for roughly 1–5% of the cost.

Screens a name against OFAC, EU, UK, and UN sanctions lists plus Wikidata-derived PEP watchlists with fuzzy matching, returning an instant JSON verdict.

Fix spelling, grammar, and punctuation in text without changing its meaning, using an LLM at a fraction of frontier model cost.

Fixes malformed JSON (truncated, mis-quoted, trailing commas, etc.) into valid, parseable output using LLM-based repair at low cost

Joins two in-memory tables by key using inner, left, right, or outer join semantics, priced per MB of input data

Verifies a JWT signature with algorithm pinning, blocking alg:none and algorithm confusion attacks, without logging the key.

Renders an HTML string to a PNG/image using a sandboxed headless Chromium instance and returns the screenshot

Renders bar, line, or pie charts from JSON data and returns a report-ready PNG image in a single API call.

Extracts structured tables from documents as rows and columns using AWS Textract, returning clean tabular data instead of raw text.

Converts Markdown text to sanitized, embed-safe HTML with whitelist-based XSS protection

Detects silent intervals in audio as JSON or trims silence out of audio files to produce tighter, cleaner recordings.

Extracts the dominant color of an image and returns it as a hex code and RGB values for use in theming and palette generation.

Computes a keyed HMAC digest (SHA-256, SHA-512, SHA-1, or MD5) from a secret key and data payload, without logging the key.

Parses invoices and receipts from uploaded files into structured fields (vendor, totals, dates, line items, SKUs) using AWS Textract AnalyzeExpense for $0.055 per document.

Extracts clean article text — title, byline, and body — from any public URL, stripping boilerplate and navigation clutter.

Answers a specific question strictly from the provided context text using a high-accuracy LLM, with no hallucinated sources and tiered model selection.

Renders a clean, branded invoice PDF from structured JSON containing line items, totals, and branding information

Extracts all regex pattern matches from a text string, with ReDoS protection and bounded inputs.

Resizes or fits images to specified dimensions, preserving animated GIF/WebP frames, priced at $0.0003 per megapixel

Ask natural-language questions against a document (PDF, image, etc.) and receive grounded, extracted answers using AWS Textract Queries — no OCR pipeline required.

Validates rows in a CSV or structured file against a user-defined JSON schema, returning row-level validation errors as JSON

Extracts text from scanned PDF pages using Tesseract OCR, making paper documents machine-readable and searchable.

Combines multiple PDF files into a single PDF in a user-specified order, billed at $0.001 per page

Renders a logic-less Mustache-subset template safely against a provided data context, with no eval and injection protection.

Converts Word (.docx/.doc), Excel (.xlsx/.xls), or PowerPoint (.pptx/.ppt) files to PDF using LibreOffice, preserving layout.

Compares two PDF documents and returns a unified diff highlighting every textual change between versions

Applies a unified diff (patch) to a source string, returning the patched result or a 422 error on context mismatch — never a silent bad merge.

Detects faces in an image and returns attributes including age range, emotions, pose, and facial landmarks using AWS Rekognition.

Cuts a specified time range from an audio or video file using fast stream-copy, preserving original codecs with no quality loss.

Applies Gaussian blur to an image at a configurable strength, suitable for background blurring, redactions, and thumbnail softening.

Generates a deterministic GitHub-style identicon PNG from any seed string, producing a stable avatar for users or agents

Detects objects, scenes, and concepts in an image using AWS Rekognition, returning labels with confidence scores

Converts tabular data between CSV, TSV, JSON, and NDJSON formats while preserving structure, billed per MB.

Extracts N evenly-sampled frames from a video and returns a manifest of presigned image links

Parses an uploaded Excel (.xlsx) file into structured CSV or JSON, handling multiple sheets, headers, and data type detection.

Renders any public web page to a PDF file using headless Chromium with SSRF-guarded egress

Extracts all embedded images from a PDF file and returns them as PNG files, billed at $0.001 per page.

Normalizes audio file loudness to broadcast-standard EBU R128 targets, adjusting integrated loudness (LUFS), true-peak, and loudness range (LRA).

Renders an HTML document to a print-quality PDF using sandboxed headless Chromium

Inspects any media file via ffprobe and returns duration, container format, streams, codecs, dimensions, and bitrate as structured JSON

Mixes multiple audio tracks into a single output using amix, with optional loudness normalization — ideal for combining voiceover and background music in one API call

Converts any image to grayscale, priced per megapixel, accepting any common input format.

Converts tabular data into styled, multi-sheet Excel (.xlsx) files with formula-injection protection

Parses a government-issued ID (passport or driver's license) into typed structured fields using AWS Textract AnalyzeID

Free-text search for places and points of interest, returning names, addresses, and coordinates as JSON

Automatically detects and crops uniform borders from images without requiring manual coordinates, at $0.0003 per megapixel.

Executes full SQL queries against CSV, JSON, or Parquet files using a sandboxed read-only DuckDB engine, returning results as Parquet

Classifies text as positive, negative, or neutral sentiment using LLM-grade accuracy at a fraction of the cost

Converts a WGS-84 longitude/latitude coordinate pair into a nearest street address via reverse geocoding.

Joins 2–10 video or audio clips end-to-end via stream-copy (no re-encode) at $0.0005/MB, billed by the megabyte with a 1¢ x402 minimum.

Reads and returns dimensions, format, and EXIF presence for an image URL before processing

Changes the playback speed of audio or video files with pitch preservation by default, billed per megabyte of input.

Delete, reorder, or insert pages within a PDF document without requiring desktop software

Converts images between PNG, JPEG, WebP, and AVIF formats (with HEIC input support), preserving animation when requested, billed per megapixel.

Reads and returns a PDF's outline/bookmark tree as structured JSON, enabling navigation of document structure before content extraction.

Hashes arbitrary bytes using SHA-256, SHA-512, SHA-1, or MD5 and returns the digest as a hex string, billed per megabyte of input.

Extracts clean text and structure from a PDF file, optimized as input for LLMs and search pipelines

Recompresses and linearizes a PDF to reduce file size and enable fast web viewing (progressive rendering)

One-call RAG pipeline: embeds a question, searches a vector memory index, and optionally returns a grounded answer with cited source chunks

One-call RAG pipeline that chunks text, embeds every chunk, and writes vectors into a durable persistent vector index — no separate vector DB required.

Adjusts brightness, saturation, hue, lightness, negation, and tint of an image in a single API call

Creates a scoped, single-use read token (burn-after-read share link) to hand a stored file to another agent or person without requiring them to have an account.

Decodes a QR code from an image and returns the embedded text or URL as JSON

Save and restore durable agent state snapshots — write a checkpoint at any point and resume from it later, billed per-use with no account required.

Extracts text from an image using Tesseract OCR, supporting 10 languages and returning plain text output

Extracts key terms from text as a clean list for tagging, SEO, and indexing using LLM inference at near-Opus-4.8 accuracy

Generates a single concise title for any text input — suitable for headlines, email subjects, or filenames — using LLM inference at low cost

Executes a server-side chain of multiple operations (e.g. OCR → extract → convert → store) in a single API call with one payment, returning a single result.

Append, read, and collaborate on a durable shared scratchpad with per-collaborator tokens; supports multi-agent workflows with persistent working memory.

Split a PDF into individual pages or defined page ranges, billed at $0.001 per page via x402 micropayments

Password-protects a PDF with 256-bit AES encryption and sets access permissions

Removes duplicate rows from a CSV dataset in a single API call, preserving original row order.

Coerces CSV/tabular column values into specified types (string, integer, number, boolean, date) to eliminate stringly-typed data problems

Sample rows from a dataset by count or fraction, using random, head, or systematic strategies with a reproducible seed

Groups rows of tabular data and computes aggregations (count, sum, avg, min, max) in a single API call, billed per MB of input data.

Read or write media container tags (e.g. ID3, EXIF, MP4 atoms) as JSON without re-encoding the file via stream-copy

Counts characters, words, and lines in text and returns estimated reading time, billed at $0.0002/MB via x402.

Batch-writes embeddings into a prepaid vector index inside a Relaystation baton (durable shared-storage object), enabling agent RAG memory without self-hosting a vector database.

Zips one or more files (from URLs or raw payloads) into a single downloadable archive and returns a download link

Slices rows from a dataset by offset, limit, head, or tail for pagination or preview use cases

Reshapes tabular data between long and wide formats — pivot or unpivot a table without a spreadsheet.

Generates realistic fake rows of structured test data from a user-defined field→type schema, returning results as JSON, CSV, or NDJSON.

Sort rows in a dataset by one or more columns, ascending or descending, billed at $0.0002/MB via x402

Generates a print-ready 1D barcode image (PNG) from a value using a specified symbology such as Code128, EAN, or UPC, with configurable height and scale.

Decodes a JWT token's header and payload without signature verification, exposing all claims instantly

Converts color values between hex, RGB, and HSL formats, returning exact converted values as JSON

Fills empty/null cells in tabular data using value, forward-fill, backward-fill, mean, median, or mode strategies

Tiles multiple images into a single contact-sheet thumbnail grid PNG for visual indexing in one API call

Crops a rectangular region from an image using exact pixel coordinates, returning the cropped image output.

Generate single or batch text embeddings at $0.002/10KB, designed to pair with the Baton vector index for pay-per-query RAG memory pipelines.

Upload a file and receive a durable, shareable public URL — no account or signup required, billed per-call via crypto wallet.

Extracts a single video frame as a PNG or JPEG image from any video at a specified timestamp

Rewrites input text in a specified style or tone (formal, plain, concise, brand voice) using LLM inference at ~1–5% of frontier model cost

Converts a video segment into a palette-optimized animated GIF suitable for sharing and looping

Generates a frame-sampled thumbnail sprite sheet from a video URL, suitable for scrubber timelines and video preview strips.

Re-encodes video or audio to a different container or codec (H.264, AAC, etc.) with optional scaling, cropping, and rotation in a single pass

Computes a unified diff of two text strings, showing exactly what changed between them

Adds computed columns to structured data using injection-safe expressions without formula evaluation

Re-encodes audio files to a different format, bitrate, or sample rate (mp3, aac, wav, flac, and more) at $0.0005/MB

Converts a free-text address string into geographic coordinates (lat/lng) and a normalized address using AWS Location Service.

Generates placeholder images at any specified size with optional label text, returning PNG or SVG output

Creates binding-by-goodwill agent-to-agent contracts with terms, parties, signatures, and paid arbitration support for $0.02 per contract

Rotates PDF pages — all pages or a selection — by any multiple of 90 degrees, billed per page at $0.001.

List or extract files embedded as attachments inside a PDF document programmatically

Renders an audio file as a waveform or spectrogram PNG image for visual track analysis

Converts text to a specified case format — uppercase, lowercase, title case, camelCase, snake_case, or kebab-case — billed per MB of input.

Sharpens a soft or blurry image with adjustable strength, returning a crisper version suitable for scans, photos, and thumbnails

Filter rows in a dataset by a structured predicate — injection-safe, no formula evaluation — billed per MB processed

Instantly profiles a dataset returning column types, null counts, cardinality, and min/max/mean statistics before you commit to full processing.

Top-k similarity search over a named baton's vector index — send a vector or raw text, get nearest matches with similarity scores

Renders social/Open Graph card PNG images with titles, colors, and branding for use in link previews

Base64-encodes or decodes a string, returning the transformed result instantly at a per-MB rate.

Converts arbitrary text into a URL-safe slug by stripping diacritics, lowercasing, and replacing spaces/punctuation with hyphens

Diagnoses and repairs damaged or corrupted PDF files, recovering files that other tools reject

Verifies a PDF's digital signatures with structural and digest checks, returning results as JSON

Generates a complete favicon set (multiple PNG sizes) from an input image URL or initials/text string

Watermarks an image or logo onto a video with configurable position and opacity control

Classifies text into one of a user-defined label set (routing, triage, moderation) using LLM-grade accuracy at a fraction of frontier model cost

OCR a scanned single-page document into a searchable PDF, extract its tables via AWS Textract, convert them to XLSX, and store both artifacts at durable URLs — all in one API call.

Re-encodes an image at lower quality to reduce file size for bandwidth-efficient delivery

Removes rows containing empty/null cells from a tabular dataset in a single API call, returning the cleaned data.

Keeps, drops, or reorders columns in a CSV table to reshape data for downstream pipeline steps.

Automatically infers a JSON Schema or SQL DDL definition from raw data, eliminating the need to hand-write schemas.

Generates one or a batch of version 4 UUIDs via a single API call, billed at $0.0001 per call with x402 micropayment support.

Rotates an image by any angle or flips it horizontally/vertically, returning the transformed image.

Renames columns in tabular data by mapping old column names to new ones, tidying schemas between pipeline steps.

Read or set PDF document metadata fields such as title, author, keywords, and dates.

Unnests (explodes) array cells in nested JSON, flattening each array element into its own row for tabular analysis

Computes a row-level changeset between two tables, returning added, removed, and changed rows

Concatenates multiple tables into a single dataset using an ordered key-union schema, safely stacking rows from heterogeneous sources.