Yogatik tool catalogue

How tool calling works here

The model is shown a ranked, capped list of tools relevant to what you asked. It replies with a call, the browser executes it locally, and the result is fed back for the next step — repeating until the model has what it needs to answer. Models that cannot do native function calling fall back to a text protocol automatically, so they keep their tools instead of losing them.

Web & research

Tool	What it does
`web_search`	Search the live web using the local desktop search sidecar (Electron only).
`deep_research`	Research a topic on the live web: runs a search, then reads the top result pages in parallel and returns their extracted text with source URLs, publication dates and authors.
`web_extract`	Read a specific web page and return its main article text, with title, author and publication date.
`browser_control`	Open and USE a real web browser the user can watch: navigate, read the page structure, click, type, scroll and manage tabs.
`browser_autopilot`	Fetch a web page's static HTML and extract its text, optionally focused on a query.
`link_preview`	Get a link preview (title, description, image) for a URL.
`youtube`	Get YouTube video information and best-effort caption transcript when available.
`wikipedia`	Look up an encyclopedia article on Wikipedia and return its summary, or search Wikipedia by topic.
`scholar`	Search academic literature (OpenAlex, plus arXiv preprints).
`stackoverflow`	Search Stack Overflow for programming problems and return the accepted or highest-voted answers.
`hackernews`	Search Hacker News discussions.
`archive`	Find an archived snapshot of a URL in the Internet Archive Wayback Machine.
`wayback_archive`	Check if a website, article, or URL has historical snapshots archived on the Internet Archive Wayback Machine.
`rss_feed`	Read RSS/Atom feed (presets: hackernews, bbc, techcrunch, reddit_tech, arxiv_ai).
`firecrawl`	Firecrawl deep web crawler and RAG Markdown dataset engine (inspired by mendableai/firecrawl).
`lightpanda`	Lightpanda ultra-fast headless browser and RAG scraper (inspired by lightpanda-io/browser).
`web_automate`	Execute multi-step web automations: batch extraction from multiple URLs in parallel, HTML table scraping into structured data, change tracking / diff monitoring, or RSS monitoring.

Documents & knowledge

Tool	What it does
`doc_search`	Search the documents the user has uploaded in this app (PDF, text, markdown, CSV, JSON).
`doc_list`	List the documents the user has uploaded, with sizes.
`local_vault_search`	Search 100% locally across all stored user documents, notes, and past conversation history in IndexedDB.
`pdf_extract`	Extract text from a PDF file.
`summarize`	Summarize a block of text into key sentences.
`doc_export`	Convert Markdown or structured text into Microsoft Word (.doc), PowerPoint presentation (.pptx), CSV spreadsheet (.csv), HTML, or formatted JSON for 1-click downloading.
`doc_enhance`	Enhance and edit generated documents (PPT presentations, Word DOCs, CSV spreadsheets, Markdown reports) by adding topic illustration images to slides, auto-generating summary data tables, calculating CSV totals/averages, or adding executive styling.
`document_generator`	Generate professional, standalone documents completely client-side in multiple open-source formats: Word documents (.doc/.docx), interactive HTML5 presentation slide decks, printable business invoices, awards/certificates, or spreadsheets.
`md_to_pdf`	Convert markdown text to a polished, downloadable PDF (headings, lists, tables, code, blockquotes).
`text_analytics`	Analyze text for readability metrics, word/sentence counts, sentiment polarity, and reading time.
`keyword_extract`	Extract the most important keywords and keyphrases from text using local frequency analysis.
`entity_extract`	Extract likely people, organizations, locations, dates, numbers, emails, and URLs from text.
`query_refine`	Refine a messy user query into a cleaner search query, a few focused subqueries, and an intent label.
`turbovec`	TurboQuant vector database and RAG search engine.
`memory`	Durable, on-device memory across sessions.

Code & data

Tool	What it does
`code_execute`	Execute Python in the browser (Pyodide WASM).
`js_execute`	Execute JavaScript in a sandboxed Web Worker (no DOM, no network to the page, no app access).
`code_format`	Clean, format, and fix indentation on source code snippets (JavaScript, HTML, CSS, JSON, Python, SQL).
`code_review_diff`	Perform Alibaba Open Code Review on git diffs or patch strings with line-level defect classification (SECURITY, BUG_RISK, PERFORMANCE, DESIGN, STYLE).
`code_review_scan`	Deep audit and scan an entire source file for security flaws, bug risks, performance pitfalls, and design flaws without needing a git diff.
`code_review_pr`	Generate an executive PR review report with overall risk score (0-100) and automated merge verdict (APPROVE, COMMENT, REQUEST_CHANGES).
`diff`	Compare two texts and show differences.
`regex`	Test, find, or replace using regular expressions.
`data_convert`	Convert between JSON and CSV (RFC 4180, Excel-safe), or prettify/minify JSON.
`data_stats`	Compute numerical statistics (mean, median, min, max, std dev, counts) on CSV or JSON datasets.
`hash`	Generate hash, encode/decode, or create UUID.
`uuid`	Generate random UUID v4 identifiers (RFC 4122).
`number_base`	Convert an integer between bases (2–36).
`numbat`	Perplexity AI Numbat agent runtime security, pre-action interception, and behavioral threat detection harness.
`repo_finder`	GitHub repository finder, similarity discovery, and open-source intelligence tool.
`fprime`	NASA F Prime (F´) flight software generator and topology validator.

Files, shell & desktop

Tool	What it does
`fs_read`	Read a UTF-8 text file inside this chat’s folders.
`fs_write`	Create or overwrite a text file in this chat’s folders.
`fs_edit`	Edit an existing text file by exact string replacement (like a targeted patch).
`fs_list`	List files and folders under a path in this chat’s folders.
`fs_search`	Search file contents across every folder bound to this chat for a substring or regex.
`fs_find_files`	Find files across the workspace by filename, glob pattern (e.g.
`fs_batch_read`	Read multiple text files in parallel inside the active workspace in a single round-trip.
`fs_batch_write`	Create or overwrite multiple files across workspace in a single round-trip.
`fs_file_tree`	Generate a compact ASCII directory tree of the workspace up to a specified depth.
`fs_file_info`	Inspect detailed metadata, file size, line counts, extension, and integrity stats for a file.
`fs_replace_content`	Replace a single contiguous block of text in an existing file.
`fs_multi_replace`	Perform multiple non-contiguous block replacements across a file in a single atomic transaction.
`fs_delete`	Delete a file or directory inside this chat’s folders.
`fs_mkdir`	Create a directory (and any missing parent directories) inside this chat’s folders.
`fs_move`	Move or rename a file or directory inside this chat’s folders.
`fs_undo`	Undo a previous file change made in this chat (write, edit, delete, move).
`fs_add_folder`	Open a native folder picker so the user grants this chat access to a folder on their computer.
`terminal_run`	Execute a terminal CLI command in this chat’s primary working folder.
`proc_start`	Start a LONG-RUNNING command in the background (dev server, watch-mode tests, a build that streams output) and return immediately with a process id.
`proc_output`	Read the output of a background process started with proc_start, and whether it is still running.
`proc_stop`	Stop a background process started with proc_start.
`proc_list`	List the background processes running for this chat.
`git_status`	Show git status for this chat’s working folder: current branch, and which files are staged, modified or untracked.
`git_log`	Show recent git commits (hash, author, date, subject) for this chat’s working folder.
`git_diff`	Show the git diff for this chat’s working folder — what has actually changed, line by line.
`watch`	Watch this chat’s working folders for changes made outside the app, or read which files have changed since the last check.
`watch_folder`	Start, stop, or list file-system watchers on a path inside the granted folder.
`clipboard_access`	Read the user’s current clipboard (text and images), read recent clipboard history, or write text to the clipboard.
`file_dialog`	Open a native file/folder picker or save dialog and return real OS paths.
`process_manager`	List running processes on the user’s computer, or terminate one by PID.
`system_state`	Report the desktop machine’s power and idle state: how long the user has been idle, and whether the device is on battery or AC power.
`screen_inspect`	Capture the desktop screen and inspect the active foreground application window (title, process, visual UI context, and OCR text).
`desktop_action`	Perform an OS-level synthetic action in the active application window or desktop (type text, trigger hotkeys, open apps/URLs, or copy to clipboard).
`computer_control`	Control the mouse and keyboard across any application on the user’s computer: click or double-click at a screen coordinate, move the pointer, scroll, or press a named key/combo (enter, tab, escape, ctrl+c, ctrl+shift+t, f5…).

Images, audio & video

Tool	What it does
`image_generate`	Generate a high-fidelity image from a text prompt with 8K details, cinematic lighting, aspect ratios, and negative prompt filtering.
`sticker_generate`	Generate a sticker, icon, badge, or graphic illustration from a text prompt.
`image_info`	Get image dimensions, size, and basic metadata.
`chart`	Generate a chart/graph from data.
`diagram`	Generate a diagram from Mermaid syntax (flowchart, sequence, class, etc.).
`diagram_render`	Generate visual Mermaid.js flowcharts, mind maps, sequence diagrams, or entity graphs.
`ocr`	Extract text from an image using OCR.
`unlimited_ocr`	Baidu Unlimited-OCR long-context document parser and LaTeX formula recognition engine (inspired by baidu/Unlimited-OCR).
`qr_generate`	Generate a QR code from text/URL.
`qr_read`	Read/decode a QR code from an image.
`video_render`	Render a real video file (MP4) in the browser from a scene list — titles, bullet slides, images with Ken Burns motion, and animated bar charts, with cross-fades and spoken narration.
`video_edit`	Edit an EXISTING video: trim to a time range, change speed, resize or re-encode at a different quality.
`audio_edit`	Get audio file information (duration, channels, sample rate).
`tts`	Speak text aloud through the device speakers.
`stt`	Record from the microphone and transcribe it.
`text_to_audio`	Narrate text into a downloadable audio file (WAV/MP3) using on-device neural voice (Kokoro) or ElevenLabs studio cloud voices.
`podcast_generate`	Generate a multi-speaker audio podcast / conversational overview (Host & Guest discussion) from a topic or script using studio ElevenLabs voices or on-device neural voices.
`audio_overview`	Generate a multi-speaker audio podcast / conversational overview (Host & Guest discussion) from a topic or script using studio ElevenLabs voices or on-device neural voices.
`see`	Look through the user's camera and answer a question about what is visible.
`color_palette`	Extract dominant colors from an image or convert color formats.
`threeui`	ThreeUI 3D web component and procedural Three.js generator (inspired by MengTo/threeui).

Agents & automation

Tool	What it does
`spawn_agents`	Delegate focused sub-tasks to specialist sub-agents that run concurrently, then get their merged results.
`crew_orchestrator`	Execute multi-agent autonomous crew workflows inspired by CrewAI, AutoGen, and LangGraph.
`auto_skills`	Manage auto-generated skills learned from successful problem-solving sessions.
`scheduler`	Manage background scheduled jobs, alarms, timers, and reminders.
`timer`	Set, list, or cancel alarms, countdown timers, and reminders in the browser.
`todo`	Track the steps of a multi-step task for this conversation.
`agent_reach`	Agent-Reach multi-platform intelligence gathering & open web reader (inspired by Panniantong/Agent-Reach).
`guardrails`	Guardrails & Rebuff safety enforcement engine (inspired by guardrails-ai and protectai/rebuff).
`semantica_record_decision`	Record an auditable, structured AI decision in the Context Graph with reasoning, category, confidence, and metadata for compliance and precedent search.
`semantica_trace_causal_chain`	Trace the causal ancestry (upstream causes) and impact tree (downstream effects) of a specific AI decision node.
`semantica_find_precedents`	Find past similar AI decisions and precedents matching a query scenario or context.
`semantica_context_graph`	Inspect, query neighbors, detect semantic conflicts, or extract entity-relation triplets in the Context Graph.
`semantica_audit_export`	Export regulator-ready audit trails of AI decisions and knowledge graphs in JSON, CSV, Markdown, or W3C PROV-O RDF/Turtle format.

Finance & markets

Tool	What it does
`finance_analytics`	Run financial valuation and risk maths exactly, on-device: DCF valuation, NPV, IRR, CAGR, and risk statistics for a price series (volatility, Sharpe, Sortino, max drawdown, Value at Risk, expected shortfall).
`market_data`	Fetch historical market prices or economic indicators without an API key.
`crypto_price`	Look up live cryptocurrency prices, 24-hour price change percentage, 24h high/low, and trading volume.
`currency`	Convert between currencies at European Central Bank reference rates, or list rates.
`world_bank`	Retrieve official macroeconomic and developmental statistics from the World Bank Open Data catalog for 218+ countries.

Everyday & open data

Tool	What it does
`weather`	Get current weather and 7-day forecast for a location.
`air_quality`	Current air quality (US + European AQI, PM2.5, PM10, ozone, NO₂, CO) and pollen for a location.
`calculator`	Evaluate a mathematical expression.
`unit_convert`	Convert between units (length, weight, volume, temperature).
`translate`	Translate text between languages.
`grammar_check`	Proofread text for grammar, spelling, punctuation and style using LanguageTool.
`geocode`	Convert a place name into coordinates and structured address data using OpenStreetMap, or reverse a coordinate pair into a place.
`solar_times`	Calculate precise astronomical and solar times for any latitude and longitude (or city location).
`earthquake`	Recent earthquakes from the USGS feed, optionally near a location.
`iss_location`	Track the real-time orbital location, latitude, longitude, altitude, velocity, and visibility of the International Space Station (ISS).
`dictionary`	Define an English word: meanings, part of speech, pronunciation, synonyms and example usage.
`thesaurus`	Find synonyms, antonyms, rhymes, or related words for a word (Datamuse, open + keyless).
`books`	Search Open Library (Internet Archive) for books by title, author or subject.
`gutenberg`	Search Project Gutenberg for public-domain books and optionally fetch the full text.
`country_info`	Look up facts about a country: capital, population, region, currencies, languages, timezones, calling code, flag.
`package_info`	Look up a software package or repository: current version, licence, dependencies, maintenance status and links.
`drug_info`	Look up official prescription and over-the-counter (OTC) drug information from the US Food & Drug Administration (FDA) database.
`chemical_info`	Look up chemical compound properties, molecular structures, and formulas from the National Institutes of Health (NIH) PubChem database.
`nobel_prize`	Search the official Nobel Foundation database for Nobel Prize winners, categories, citations, and laureates.
`trivia_quiz`	Generate verified trivia questions and answers across diverse categories (Science, Computers, Mathematics, History, Geography, Books, Film, General Knowledge).
`postal_lookup`	Look up geographic location, city, state, province, latitude, and longitude for any postal code or ZIP code worldwide across 60+ countries.
`nasa_apod`	Fetch NASA Astronomy Picture of the Day (APOD).
`nasa_asteroids`	Track near-Earth asteroids and comets using NASA Near Earth Object Web Service (NeoWs).
`itunes_search`	Search Apple iTunes catalog for songs, music tracks, albums, artists, podcasts, and audiobooks.
`art_institute`	Search 100,000+ world-class artworks from the Art Institute of Chicago collection.
`dns_lookup`	Perform browser-native DNS lookups via Google and Cloudflare DNS-over-HTTPS.
`activity_suggest`	Find curated activities and ideas to overcome boredom, learn a new hobby, or plan social events.
`jokes`	Generate funny programming, developer, and general humor jokes with setups and punchlines.
`animal_facts`	Get verified feline & canine animal facts, cat trivia, and dog breed intelligence.
`cocktail_recipe`	Search drink recipes, cocktails, and non-alcoholic mocktails from TheCocktailDB.
`federal_register`	Search official US Government executive orders, presidential proclamations, federal agency regulations, and public rules from the Federal Register API.
`bible_scripture`	Look up Bible verses, passages, and chapters across multiple public translations (World English Bible, King James Version, etc.) via the Bible API.
`wikimedia_feed`	Retrieve Wikipedia official featured content feed for today or any date: Today in History ("On this day" historical anniversaries), today's Featured Article, and Picture of the Day.
`user_profile_gen`	Generate realistic mock user profiles, testing personas, and synthetic identity data (name, email, street address, city, country, phone, age, username, avatar photo).
`whois`	Lookup WHOIS information for a domain.
`ip_lookup`	Lookup geolocation data for an IP address.

Security research

Tool	What it does
`deepsec`	Run Deepsec — the 5-stage agent-powered security audit harness (Scan, Investigate, Revalidate, Enrich, Export) to detect logic vulnerabilities, injection flaws, SSRF, path traversal, RCE, broken access control, and hardcoded secrets with remediations.
`hexstrike_recon`	Perform defensive reconnaissance and technology fingerprinting on a target domain, URL, or endpoint.
`hexstrike_audit_headers`	Audit HTTP response headers and CORS configuration against industry security benchmarks (OWASP, NIST).
`hexstrike_vuln_scan`	Scan code, endpoint configurations, or architecture descriptions for OWASP Top 10 security weaknesses, generating CVSS v3.1 metrics and exact defensive fixes.
`hexstrike_attack_surface`	Map and categorize application attack surfaces (public APIs, authenticated admin routes, file uploaders, third-party integrations) with exposure indices.
`hexstrike_generate_playbook`	Generate an executive security report and tactical engineering remediation playbook with exact code patches and incident response steps.
`cloudflare_os`	Cloudflare OS workspace, Gatekeeper capability security, and sandboxed Gadget runtime (inspired by cloudflare/cloudflare-os).

Social & careers

Tool	What it does
`social_search`	Search live posts, tweets, threads, discussions, and creators across social media platforms.
`social_post_generator`	Format and structure high-converting social media posts and scripts.
`job_search`	Search job openings, career opportunities, and hiring posts across Naukri, Indeed, LinkedIn Jobs, and Glassdoor.

Connectors

Tool	What it does
`mcp_resource`	Access data published by connected MCP servers as "resources" (documents, records, files the server exposes).
`mcp_prompt`	Use prompt templates published by connected MCP servers.

Tool	What it does
`web_automation`	Execute multi-step web automations: batch extraction from multiple URLs in parallel, HTML table scraping into structured data, change tracking / diff monitoring, or RSS monitoring.
`password_generate`	Generate a strong random password or passphrase with a rough entropy/strength estimate.
`cron_next`	Parse a 5-field cron expression (min hour day-of-month month day-of-week) and list its next run times.
`timezone`	Convert a time between IANA time zones, or show the current time in one or more zones (world clock).
`alarm`	Set, list, or cancel alarms, countdown timers, and reminders in the browser.
`sub_agent_runner`	Manage isolated sub-agents that run in separate processes with their own terminals and Python RPC servers.
`tv_shows`	Search television series and streaming show metadata from TVMaze.

Try any of these now

No account. Add a free API key and the whole catalogue is live.

Open the web app Read the guide

Every tool, and what it is for