Skill graph

Topological identity verification, Sybil-check, zero-trust attestation, and UCAN-style delegation validation

Request and verify on-chain payments using the x402 micropayment protocol.

Fetch a single article by slug — full intro, body, FAQ, references, embedded reviewers + authors with credentials, and pre-formatted citatio…

Identity, audience, focus, sponsor relationship, crisis routing, and links for Psychiatry for Children. Always safe to call when the agent n…

Paginated list of all native articles on this microsite (clinician-reviewed). Returns lightweight summaries — call get_article for full body…

Google search results as structured JSON. Returns titles, URLs, snippets, and metadata.

Returns the canonical crisis-resource payload (911, 988 Suicide & Crisis Lifeline, Crisis Text Line). Call any time the user mentions self-h…

Full-text search of clinician-reviewed pediatric psychiatry articles published on Psychiatry for Children, ranked by relevance. Use to find …

Return formatted citation strings (AMA, APA, Chicago) for an article slug. Useful when an agent needs a verifiable source line.

Accept or initiate USDC/stablecoin micropayments via the HTTP 402 / x402 protocol for autonomous agent-to-agent value transfer.

Accept payments via Google's Agent Payments Protocol (AP2) extensions to A2A, including PayPal, Mastercard Agent Pay, and crypto x402 settle…

Execute cryptocurrency payments, transfers, and settlements on public blockchains.

Estimate current Bitcoin transaction fees, mempool pressure, send-or-wait recommendations. Covers fee rate forecasting, confirmation time pr…

Generate, send, and track invoices, handle billing cycles, accounts receivable, and payment reminders.

Construct, sign, broadcast or analyze Bitcoin transactions and wallets, including UTXO management and address analysis.

Generate buy/sell signals, technical indicators, and strategy recommendations across equity, crypto, or commodity markets.

Analyze open-interest, funding rates, multi-venue derivatives signals, regime detection, and cross-venue confluence for crypto perpetuals/op…

Analyze and trade prediction markets like Kalshi or Polymarket: edge identification, position management, settlement tracking.

Run evals on language/vision models — accuracy, bias, latency, cost — using LangSmith, OpenAI Evals, custom rubrics.

Serve ML model predictions over HTTP/gRPC with batching, autoscaling, and GPU acceleration.

Fine-tune language or vision models on custom datasets; manage runs, hyperparameters, and checkpoints.

Build, clean, label, and version training datasets; deduplicate; balance class distributions.

Generate API docs, README files, architecture diagrams, code comments from source code or specifications.

Find outdated or vulnerable dependencies; suggest safe upgrades; manage lockfiles across npm, pip, cargo, gem.

Diagnose runtime errors, stack traces, performance regressions; suggest fixes and verify with tests.

Review code changes for correctness, style, security; suggest improvements and approve/request changes on GitHub/GitLab PRs.

Search across code repositories for symbols, patterns, references; navigate cross-file relationships.

Generate new features, refactor existing modules, write tests, and integrate with frameworks across multiple languages.

Write unit, integration, and end-to-end tests; identify uncovered paths; generate fixtures and mocks.

Fetch and parse web pages, follow links, extract structured records; handle JS-rendered sites and rate limits.

Query blockchain data via The Graph subgraphs, GraphQL endpoints, or REST APIs; return tabular onchain records like swaps, balances, positio…

Call third-party REST APIs with auth, pagination, retries, and schema transformation.

Aggregate, summarize, and analyze real-time news from multiple sources, with optional regional, topical, or sentiment filtering.

Query SPARQL endpoints, Neo4j, or property-graph databases; traverse relationships and aggregate.

Embed text/code/images; perform similarity search against vector stores — Pinecone, Weaviate, Qdrant, pgvector.

Pay-per-query SQL access to multiple authoritative datasets, returning actual database records across biomedical, financial, security, and r…

Issue SQL queries across multiple databases — Postgres, MySQL, BigQuery, Snowflake — and join results.

Access regional or national government open data sources, public records, statistics, and civic datasets.

Build and operate extract/transform/load pipelines — Airflow, dbt, Fivetran, Dagster — across source systems.

Query GraphQL APIs with introspection, fragment composition, and persisted queries.

Query and manage cloud data warehouses (BigQuery, Snowflake, Redshift, Databricks SQL); transform and load.

Synthesise speech with voice cloning — ElevenLabs, Play.ht, Azure, OpenAI TTS.

Generate images from prompts via Stable Diffusion, DALL-E, Imagen, Flux; edit, inpaint, upscale.

Generate or edit video — Runway, Sora, Pika, Veo; clip splicing, captions, b-roll synthesis.

Read, modify, and generate Figma frames; export assets; sync design tokens.

Track campaign performance, calculate CAC/LTV, run multi-touch attribution across funnels.

Audit pages for SEO, generate optimised titles/descriptions, monitor rankings, find content gaps.

Create and tune ad campaigns on Google Ads, Meta, LinkedIn, TikTok with budget pacing and creative iteration.

Plan, schedule, and publish content across blogs, social, newsletters with editorial governance.

Managing real-time collaboration and state synchronization among agents.

Tracking latency, throughput, resource utilization, and service reliability over time.

Coordinating plans across multiple agents, resolving dependencies and optimizing sequencing.

Breaking complex objectives into structured, atomic subtasks.

Allocating responsibilities to agents based on capabilities and task requirements.

Interpreting and explaining API specifications, endpoints, parameters, and expected payloads.

Evaluating processes or outputs against defined standards (e.g., GDPR, HIPAA) and identifying gaps.

Evaluating datasets for completeness, validity, consistency, and timeliness.

Additional metadata associated with the skill.

Evaluating data handling or user flows to surface potential privacy risks and recommend mitigations.

Facilitating negotiation, conflict handling, and consensus-building between agents.

Condensing system event or transaction logs into human-readable compliance or oversight summaries.

Scanning artifacts (code, logs, documents) to identify exposed credentials, tokens, or other sensitive secrets.

Identifying indicators of malicious activity, suspicious patterns, or emerging threats across logs and data sources.

The id as a unique identifier of the skill.

Creating new images from learned patterns or data using machine learning models.

Configuring and interpreting telemetry signals, thresholds, and alerts for operational health.

Predicting numerical values based on tabular attributes and features.

Assessing outputs for accuracy, relevance, coherence, safety, and style adherence.

Reviewing code, configurations, or dependency manifests to surface potential security weaknesses and misconfigurations.

Running standardized benchmarks or evaluation suites and summarizing results.

Defining or explaining steps to allocate and configure compute, storage, and networking resources.

The name as a unique identifier of the skill.

Designing or explaining multi-step sequences that extract, transform, and load datasets.

Categorizing potential operational or data-related risks by impact and likelihood for prioritization.

Creating targeted test inputs or scenarios to probe system behavior and edge cases.

Identifying unusual patterns, drifts, or deviations in data or model outputs.

Assigning labels or categories to images based on their visual content.

Formulating high-level multi-phase strategies aligned with long-term objectives.

Organizing intermediate reasoning steps into clear, justifiable sequences.

Detecting and correcting errors, inconsistencies, and missing values to improve dataset quality.

Coordinating multi-stage application or model deployments, rollbacks, and version transitions.

Proposing plausible explanations or solution pathways for incomplete or uncertain scenarios.

Identifying and isolating key characteristics or patterns from an image to aid in tasks like classification or recognition.

Assigning labels or categories to images based on their visual content.

Selecting and ordering tool invocations to accomplish a specified goal efficiently.

Assigning labels or classes to audio content based on its characteristics.

Constructing informative transformed variables to improve downstream model performance.

Maintaining coherent reasoning chains over extended sequences of steps or time.

Capabilities for transforming content between different modalities (text, image, audio, video, or 3D).

Deriving structural metadata (fields, types, relationships) from raw or semi-structured data.

Classifying data based on attributes using classical machine learning approaches.

Assigning labels or categories to entire videos or segments based on their visual and audio content.

Designing or describing automated sequences integrating multiple tools or services.

Converting between any supported modalities (text, image, audio, video, or 3D).

Transforming audio through various manipulations including cutting, filtering, and mixing.

Designing or modifying continuous integration and delivery workflows and pipelines.

Identifying and locating specific objects within an image or video, often by drawing bounding boxes around them.

Translating organizational or regulatory policies into structured, enforceable rules or checklists.

The base skill is a generic skill defining a set of attributes for skill classes; not meant to be used as an actual skill.

Predicting the distance or depth of objects within a scene from a single image or multiple images.

Document or database question answering is the process of retrieving and using information from a document or database to answer a specific …

Generation of any is augmenting the creation of text, images, audio, or other media by incorporating retrieved information to improve or gui…

The process of converting a 2D image into a 3D representation or model, often by inferring depth and spatial relationships.

Transforming one image into another using a learned mapping, often for tasks like style transfer, colorization, or image enhancement.

Identifying and locating specific points of interest within an image or object.

Producing segmented regions in an image to highlight specific areas or objects, typically represented as separate layers or overlays.

Tracking, promoting, and documenting different iterations of models and their artifacts.

Linking custom scripts or functions with external tools to extend capabilities.

Compute routes, ETAs, distance matrices, geocoding via Google/Mapbox/HERE/OSM.

Retrieve current conditions, multi-day forecasts, and severe-weather alerts by coordinate or postal code.

Search code by intent rather than tokens — Sourcegraph, GitHub Code Search, Codeium.

Query news from GDELT, NewsAPI, Reuters; filter by source, recency, region, and topic.

Web search via Brave, Tavily, Serper, Bing; return ranked results with snippets and source URLs.

Search and retrieve scientific papers — arXiv, Semantic Scholar, PubMed, OpenAlex — with full metadata.

Parse resumes, score against job requirements, identify red/green flags, schedule next steps.

Source candidates from LinkedIn, GitHub, AngelList; build ICP filters; draft outreach.

Compute payroll, handle taxes/benefits/garnishments, integrate with Gusto, Rippling, ADP, Deel.

Expose a merchant's product/service catalog via Google's Universal Commerce Protocol with Schema.org markup so AI shopping agents can discov…

Help users find or compare products: search across catalogs, recommendations, price comparison, where-to-buy routing.

Search and identify replacement parts, compatibility checks, supplier sourcing across catalogs.

Surface early-stage startup signals: founder activity, GitHub momentum, funding indicators for deal sourcing.

Moving company quotes by inventory and distance, schedule the move, packing add-ons, insurance options.

HVAC maintenance plans, urgent service calls, seasonal tune-ups, system replacement quotes.

Electrical-service intake, quote ranges by job type, scheduled site visits, panel/wiring/installation work.

Emergency plumber dispatch, rough-estimate quotes by problem type, scheduled service for renovations.

One-time and recurring home/office cleaning bookings, customization (deep clean, move-in/out), team dispatch.

Lawn mowing, landscaping design, seasonal cleanups, snow removal scheduling for residential properties.

Verify identities, run know-your-customer flows, sanctions screening, and risk classification for regulated workflows.

Review commercial or professional contracts for risks, missing clauses, regulatory issues, and negotiation points with jurisdiction awarenes…

Analyse public equities, ETFs, real estate; build watchlists; backtest strategies.

Build budgets, run scenario analysis, forecast revenue and cash flow with sensitivity inputs.

Record transactions, reconcile bank feeds, manage chart of accounts — QuickBooks, Xero, NetSuite.

Compute tax liability, file VAT/sales tax/withholding, generate tax reports across jurisdictions.

An agent pays a price to a participant.

An agent orders a (not yet released) object/product/service to be delivered/sent.

The act of registering to be a user of a service, product or web page.\n\nRelated actions:\n\n* [[JoinAction]]: Unlike JoinAction, RegisterA…

An agent inspects, determines, investigates, inquires, or examines an object's accuracy, quality, condition, or state.

An agent joins an event/group with participants/friends at a location.\n\nRelated actions:\n\n* [[RegisterAction]]: Unlike RegisterAction, J…

The act of notifying someone that a future event/action is going to happen as expected.\n\nRelated actions:\n\n* [[CancelAction]]: The anton…

The act of editing a recipient by removing one of its objects.

The act of forming a personal connection with someone/something (object) unidirectionally/asymmetrically to get updates polled from.\n\nRela…

This is the [[Action]] of navigating to a specific [[startOffset]] timestamp within a [[VideoObject]], typically represented with a URL temp…

The action of authenticating into a device or application.

An agent controls a device or application.

The act of deliberately creating/producing/generating/building a result out of the agent.

The act of notifying someone of information pertinent to them, with no expectation of a response.

The act of inserting at the beginning if an ordered collection.

The act of resuming a device or application which was formerly paused (e.g. resume music playback or resume a timer).

An agent tracks an object for updates.\n\nRelated actions:\n\n* [[FollowAction]]: Unlike FollowAction, TrackAction refers to the interest on…

The act of transferring/moving (abstract or concrete) animate or inanimate objects from one place to another.

The act of managing by changing/editing the state of the object.

The act of starting or activating a device or application (e.g. starting a timer or turning on a flashlight).

The act of organizing tasks/objects/events by associating resources to it.

An agent bookmarks/flags/labels/tags/marks an object.

The act of conveying information to another person via a communication medium (instrument) such as speech, email, or telephone conversation.

The act of expressing a difference of opinion with the object. An agent disagrees to/about an object (a proposition, topic or theme) with pa…

An agent approves/certifies/likes/supports/sanctions an object.

An agent leaves an event / group with participants/friends at a location.\n\nRelated actions:\n\n* [[JoinAction]]: The antonym of LeaveActio…

The act of transferring money from one place to another place. This may occur electronically or physically.

The act of playing a video game.

An agent quotes/estimates/appraises an object/product/service with a price at a location/store.

The act of consuming written content.

The act of giving money in return for temporary use, but not ownership, of an object such as a vehicle or property. For example, an agent re…

The act of producing a balanced opinion about the object for an audience. An agent reviews an object with participants resulting in a review…

Scheduling future actions, events, or tasks.\n\nRelated actions:\n\n* [[ReserveAction]]: Unlike ReserveAction, ScheduleAction allocates futu…

The action that takes in a math expression and directs users to a page potentially capable of solving/simplifying that expression.

The act of forming a personal connection with someone/something (object) unidirectionally/asymmetrically to get updates pushed to.\n\nRelate…

The act of participating in an exchange of goods and services for monetary compensation. An agent trades an object, product or service with …

The act of expressing a preference from a fixed/finite/structured set of choices/options.

The act of committing to/adopting an object.\n\nRelated actions:\n\n* [[RejectAction]]: The antonym of AcceptAction.

The act of accomplishing something via previous efforts. It is an instantaneous action rather than an ongoing process.

The act of editing by adding an object to a collection.

The act of expressing a consistency of opinion with the object. An agent agrees to/about an object (a proposition, topic or theme) with part…

The act of inserting at the end if an ordered collection.

The act of registering to an organization/service without the guarantee to receive it.\n\nRelated actions:\n\n* [[RegisterAction]]: Unlike R…

The act of arriving at a place. An agent arrives at a destination from a fromLocation, optionally with participants.

The act of posing a question / favor to someone.\n\nRelated actions:\n\n* [[ReplyAction]]: Appears generally as a response to AskAction.

The act of forming one's opinion, reaction or sentiment.

The act of allocating an action/event/task to some destination (someone or something).

The act of granting permission to an object.

The act of forming a personal connection with someone (object) mutually/bidirectionally/symmetrically.\n\nRelated actions:\n\n* [[FollowActi…

The act of obtaining an object under an agreement to return it at a later date. Reciprocal of LendAction.\n\nRelated actions:\n\n* [[LendAct…

The act of giving money to a seller in exchange for goods or services rendered. An agent buys an object, product, or service from a seller f…

The act of asserting that a future event/action is no longer going to happen.\n\nRelated actions:\n\n* [[ConfirmAction]]: The antonym of Can…

The act of an agent communicating (service provider, social media, etc) their arrival by registering/confirming for a previously reserved se…

The act of an agent communicating (service provider, social media, etc) their departure of a previously reserved service (e.g. flight check-…

The act of expressing a preference from a set of options or a large or unbounded set of choices/options.

The act of generating a comment about a subject.

The act of ingesting information/resources/food.

The act of producing/preparing food.

The act of stopping or deactivating a device or application (e.g. stopping a timer or turning off a flashlight).

The act of departing from a place. An agent departs from a fromLocation for a destination, optionally with participants.

The act of discovering/finding an object.

The act of expressing a negative sentiment about the object. An agent dislikes an object (a proposition, topic or theme) with participants.

The act of providing goods, services, or money without compensation, often for philanthropic reasons.

The act of downloading an object.

The act of producing a visual/graphical representation of an object, typically with a pen/pencil and paper as instruments.

The act of swallowing liquids.

The act of swallowing solid objects.

The act of participating in exertive activity for the purposes of improving health and fitness.

The act of capturing sound and moving images on film, video, or digitally.

The act of finding an object.\n\nRelated actions:\n\n* [[SearchAction]]: FindAction is generally lead by a SearchAction, but not necessarily…

The act of transferring ownership of an object to a destination. Reciprocal of TakeAction.\n\nRelated actions:\n\n* [[TakeAction]]: Reciproc…

The act of intentionally disregarding the object. An agent ignores an object.

The act of adding at a specific location in an ordered collection.

The act of installing an application.

The act of interacting with another person or organization.

The act of asking someone to attend an event. Reciprocal of RsvpAction.

The act of providing an object under an agreement that it will be returned at a later date. Reciprocal of BorrowAction.\n\nRelated actions:\…

The act of expressing a positive sentiment about the object. An agent likes an object (a proposition, topic or theme) with participants.

The act of consuming audio content.

The action of logging into a device or application.

The act of being defeated in a competitive activity.

The act of marrying a person.

The act of an agent relocating to a place.\n\nRelated actions:\n\n* [[TransferAction]]: Unlike TransferAction, the subject of the move is a …

An agent orders an object/product/service to be delivered/sent.

The act of manipulating/administering/supervising/controlling one or more objects.

The act of producing a painting, typically with paint and canvas as instruments.

The act of participating in performance arts.

The act of capturing still images of objects using a camera.

The act of planning the execution of an event/task/action/reservation/plan to a future date.

The act of playing/exercising/training/performing for enjoyment, leisure, recreation, competition or exercise.\n\nRelated actions:\n\n* [[Li…

The act of responding instinctively and emotionally to an object, expressing a sentiment.

The act of physically/electronically taking delivery of an object that has been transferred from an origin to a destination. Reciprocal of S…

The act of rejecting to/adopting an object.\n\nRelated actions:\n\n* [[AcceptAction]]: The antonym of RejectAction.

The act of editing a recipient by replacing an old object with a new object.

The act of responding to a question/message asked/sent by the object. Related to [[AskAction]].\n\nRelated actions:\n\n* [[AskAction]]: Appe…

Reserving a concrete object.\n\nRelated actions:\n\n* [[ScheduleAction]]: Unlike ScheduleAction, ReserveAction reserves concrete objects (e.…

The action of resetting the password of a device or application.

The act of returning to the origin that which was previously received (concrete objects) or taken (ownership).

The act of notifying an event organizer as to whether you expect to attend the event.

The act of searching for an object.\n\nRelated actions:\n\n* [[FindAction]]: SearchAction generally leads to a FindAction, but not necessari…

The act of taking money from a buyer in exchange for goods or services rendered. An agent sells an object, product, or service to a buyer fo…

The act of physically/electronically dispatching an object for transfer from an origin to a destination. Related actions:\n\n* [[ReceiveActi…

The act of distributing content to people for their amusement or edification.

The act of momentarily pausing a device or application (e.g. pause music playback or pause a timer).

The act of gaining ownership of an object from an origin. Reciprocal of GiveAction.\n\nRelated actions:\n\n* [[GiveAction]]: The reciprocal …

The act of reaching a draw in a competitive activity.

The act of giving money voluntarily to a beneficiary in recognition of services rendered.

The act of traveling from a fromLocation to a destination by a specified mode of transport, optionally with participants.

The act of un-registering from a service.\n\nRelated actions:\n\n* [[RegisterAction]]: antonym of UnRegisterAction.\n* [[LeaveAction]]: Unli…

The act of applying an object to its intended purpose.

The act of consuming static visual content.

The act of expressing a desire about the object. An agent wants an object.

The act of consuming dynamic/moving visual content.

The act of dressing oneself in clothing.

The act of achieving victory in a competitive activity.

The act of authoring written creative content.

Generate sales proposals, statements of work, pricing tables tailored to deal context.

Qualify inbound leads using BANT/MEDDIC frameworks; score and route to AE.

Build prospect lists, draft personalised cold emails, follow-up sequences, and CRM tracking.

Scan for vulnerabilities (Snyk/Trivy), review IAM policies, find leaked secrets, check supply-chain risk on dependencies.

Collect, query, and alert on metrics, traces, and logs across distributed systems. Prometheus, Grafana, OpenTelemetry, Datadog, New Relic.

Author Terraform / Pulumi / CDK modules; review plans; manage state and drift.

Build, test, and deploy software via continuous integration pipelines, container orchestration, infrastructure-as-code, and release manageme…

Provision, configure, and manage AWS resources — EC2, S3, RDS, Lambda, IAM. Cost optimisation and security posture review.

Detect, triage, escalate, and resolve production incidents. Page on-call, run runbooks, write post-mortems.

Execute shell commands, manage processes, inspect filesystems, and orchestrate scripts on remote hosts.

Manage Microsoft Azure resources — VMs, App Services, Cosmos DB, AKS, Entra ID.

Provision and manage Google Cloud resources — Compute Engine, Cloud Run, BigQuery, GKE, IAM.

Operate Kubernetes clusters — deployment, scaling, troubleshooting, security policies, custom resources.

Schedule meetings on Google Calendar, Outlook, or iCal; find common availability; send invites.

Post messages, react, manage channels, and respond to events in Slack workspaces.

Compose and send transactional or bulk email via Resend/SendGrid/SES; manage threads, replies, and signatures.

Send SMS, place calls, and run interactive voice response over Twilio / Vonage / Plivo.

Read and update CRM records (Salesforce, HubSpot, Pipedrive) — contacts, deals, opportunities, activities.

Create, update, route, and resolve tickets in Zendesk, Intercom, Freshdesk, Linear, Jira Service Management.

Operate a Discord bot — slash commands, message responses, voice channels, role management.

Analyze network properties, censorship detection, anomaly detection, distributed system observability.

Facilitate introductions, welcome new members, maintain shared community context and memory across agent collectives.

Structured multi-step reasoning over knowledge graphs, planning, decomposition, and verification of complex queries.

Summarise long documents, meetings, threads into structured highlights with citations.

Translate text across languages preserving meaning, tone, and domain terminology.

Convert audio/video to text via Whisper, Deepgram, AssemblyAI; speaker diarisation and timestamps.

Edit text for grammar, clarity, voice, style guide compliance; track changes and suggestions.

Draft long-form articles, technical posts, marketing copy with research, citations, and tone matching.

Counterparty exposure checks against sanctions lists — OFAC SDN, EU consolidated, UN, UK HMT, OpenSanctions, OFAC EO 14114 (Russia secondary…

End-to-end memo validation for strategic-risk and policy briefs: evidence discipline, source coverage, sanctions-exposure flags, regulatory-…

Quote and source verification for strategic-risk memos and policy briefs. Checks that cited quotes appear verbatim in the cited source, that…

Gap analysis on which authoritative sources have (and have not) been consulted in a policy or sanctions memo — e.g. did the analyst check Op…

Logistics and trade-corridor risk assessment — Middle Corridor (Trans-Caspian), Northern Sea Route, Suez, Bab el-Mandeb, Strait of Hormuz, P…

Search and book flights across global carriers and OTAs with multi-city and flexible-date options.

Search property listings (Zillow, Funda, Rightmove, Idealista) with filters, mortgage estimates, comparable sales.

Search and book hotels, vacation rentals, and short-term stays across providers.

Access public-sector datasets — census, judiciary, parliament, statistics offices, transit — across countries.

Build and run multi-step workflows across SaaS apps — Zapier, n8n, Make, Pipedream.

Manage files in Google Drive, Dropbox, OneDrive, S3 — upload, share, version, search.

Schedule recurring jobs — cron, Temporal, Inngest, Trigger.dev — with retries and durable state.

Clone, branch, commit, push, pull-request, tag across GitHub/GitLab/Bitbucket repos.

Search and update internal knowledge bases — Notion, Confluence, Slab, Coda.

Run data protection impact assessments, document data flows, manage GDPR/CCPA/HIPAA compliance.

Triage security alerts, investigate suspicious activity, correlate signals across logs and endpoints.

Reconnaissance for authorised pentests — subdomain enumeration, port scanning, tech fingerprinting.

Text-to-3D models take in text input and produce 3D output.

Image-to-3D models take in image input and produce 3D output.

Translation is the task of converting text from one language to another.

Image to text models output a text from a given image. Image captioning or optical character recognition can be considered as the most commo…

Generating text is the task of generating new text given another text. These models can, for example, fill in incomplete text or paraphrase.

Image-to-image is the task of transforming an input image through a variety of possible manipulations and enhancements, such as super-resolu…

Any-to-any models can understand two or more modalities and output two or more modalities.

Image-text-to-image models take an image and a text prompt as input and generate a new image based on the reference image and text instructi…

Image-text-to-text models take in an image and text prompt and output text. These models are also called vision-language models, or VLMs. Th…

Question Answering models can retrieve the answer to a question from a given text, which is useful for searching for an answer in a document…

Text-to-video models can be used in any application that requires generating consistent sequence of images from text.

Visual Question Answering is the task of answering open-ended questions based on an image. They output natural language responses to natural…

Automatic Speech Recognition (ASR), also known as Speech to Text (STT), is the task of transcribing a given audio to text. It has many appli…

Document Question Answering (also known as Document Visual Question Answering) is the task of answering questions on document images. Docume…

Feature extraction is the task of extracting features learnt in a model.

Image feature extraction is the task of extracting features learnt in a computer vision model.

Sentence Similarity is the task of determining how similar two texts are. Sentence similarity models convert input texts into vectors (embed…

Summarization is the task of producing a shorter version of a document while preserving its important information. Some models can extract t…

Table Question Answering (Table QA) is the answering a question about an information on a given table.

Text Classification is the task of assigning a label or class to a given text. Some use cases are sentiment analysis, natural language infer…

Text-to-image is the task of generating images from input text. These pipelines can also be used to modify and edit images based on text pro…

Token classification is a natural language understanding task in which a label is assigned to some tokens in a text. Some popular token clas…

Video-text-to-text models take in a video and a text prompt and output text. These models are also called video-language models.

Audio classification is the task of assigning a label or class to a given audio. It can be used for recognizing which command a user is givi…

Audio-text-to-text models take both an audio clip and a text prompt as input, and generate natural language text as output. These models can…

Audio-to-Audio is a family of tasks in which the input is an audio and the output is one or multiple generated audios. Some example tasks ar…

Depth estimation is the task of predicting depth of the objects present in an image.

Masked language modeling is the task of masking some of the words in a sentence and predicting which words should replace those masks. These…

Image classification is the task of assigning a label or class to an entire image. Images are expected to have only one class for each image…

Image Segmentation divides an image into segments where each pixel in the image is mapped to an object. This task has multiple variants such…

Image-text-to-video models take an reference image and a text instructions as and generate a video based on them. These models are useful fo…

Image-to-video models take a still image as input and generate a video. These models can be guided by text prompts to influence the content …

Keypoint detection is the task of identifying meaningful distinctive points or features in an image.

Mask generation is the task of generating masks that identify a specific object or region of interest in a given image. Masks are often used…

Object Detection models allow users to identify objects of certain defined classes. Object detection models receive an image as input and ou…

Reinforcement learning is the computational approach of learning from action by interacting with an environment through trial and error and …

Tabular classification is the task of classifying a target category (a group) based on set of attributes.

Tabular regression is the task of predicting a numerical value given a set of attributes.

Text Ranking is the task of ranking a set of texts based on their relevance to a query. Text ranking models are trained on large datasets of…

Text-to-Speech (TTS) is the task of generating natural sounding speech given text input. TTS models can be extended to have a single model t…

Unconditional image generation is the task of generating images with no condition in any context (like a prompt text or another image). Once…

Video classification is the task of assigning a label or class to an entire video. Videos are expected to have only one class for each video…

Video-to-video models take one or more videos as input and generate new videos as output. They can enhance quality, interpolate frames, modi…

Visual document retrieval is the task of searching for relevant image-based documents, such as PDFs. These models take a text query and mult…

Zero-shot text classification is a task in natural language processing where a model is trained on a set of labeled examples but is then abl…

Zero-shot image classification is the task of classifying previously unseen classes during training of a model.

Zero-shot object detection is a computer vision task to detect objects and their classes in images, without any prior training or knowledge …

Route, filter, transform, or proxy data between agents; agent-native API mesh; cross-agent context layer.

Compose, orchestrate, or proxy multiple upstream APIs into agent-callable endpoints with auth, rate limit, and caching.

Provide locale-specific primitives: date conversion, address formatting, invoice formats, tax calculation, regional compliance.

Store and recall facts, preferences, and conversation history across sessions — Mem0, Letta, custom vector stores.

Capture timestamped events and decisions; reconstruct timeline; reason about temporal relationships.

Match prospects to financial advisors by focus area, schedule discovery calls, pre-collect goals and financial snapshot.

Intake buyers/sellers, surface listings, schedule viewings, mortgage pre-approval handoff.

Take legal intake by matter type (family, immigration, estate, business), book a consultation, jurisdiction and conflict-of-interest checks.

Intake for bookkeeping, tax prep, payroll for small businesses; surface engagement scope and pricing.

Match students to tutors by subject and level, schedule recurring sessions, manage packages.

Manage hotel booking inquiries, availability checks, rate quotes, room selection, and confirmation across direct and indirect channels.

Restaurant reservations, menu inquiries, dietary accommodations, room service, and food and beverage operations.

Plan, coordinate, and execute corporate meetings, social events, conferences within a venue, handling logistics and client requirements.

Manage tee time bookings, golf shop sales, lesson scheduling, and player inquiries at golf resorts and clubs.

Walk-in queue status and appointment booking for barbershops, with barber preference and service-time estimates.

Book hair appointments, surface stylists' specialties and availability, manage cuts/colors/treatments, deposit and cancellation rules.

Book manicure/pedicure/nail art services with service-duration awareness, technician preference, and add-on selection.

Spa treatment booking, modality selection (deep tissue, prenatal, etc.), package upsell, intake form pre-collection.

Match clients to tattoo artists by style, schedule consultations, manage deposits and design briefs.

Take repair-shop intake by symptom/vehicle, produce rough quotes, schedule diagnostics or service appointments.

Wash/detail package selection, scheduling, mobile-detailing dispatch, recurring subscription handling.

Tire selection by vehicle, installation appointment, seasonal storage, and alignment add-on.

Maintain authenticated browser sessions across calls — cookies, local storage, MFA — for operating on logged-in apps.

Drive browsers (Playwright, Puppeteer, Selenium) for testing, scraping, or task automation against real web apps.

Triage symptoms, schedule physiotherapy/PT appointments, pre-collect injury history, manage referrals.

Match clients to therapists by modality/insurance/availability, conduct intake, schedule first session for solo or group practices.

Book vet visits, vaccinations, surgeries; surface pet records, urgent triage, and clinic-specific protocols.

Intake and goal-setting for nutrition coaching, meal-plan options, recurring check-in scheduling, insurance/reimbursement guidance.

Schedule cleanings, exams, urgent appointments at a dental practice; surface insurance acceptance and new-patient flow.

Manage ad campaigns, bidding strategies, audience targeting, creative optimization across Google Ads, Meta, and other ad platforms.

Sense and respond to customer signals, anomalies, churn risk, and operational friction across B2B SaaS workflows.

Handle inbound inquiries for a small business: FAQ answering, appointment booking, intake routing, after-hours coverage.

Book DJs or live performers for events, song-list intake, equipment requirements, set-time scheduling.

Surface venue availability, capacity and amenities, schedule tours and walk-throughs, manage holds and contracts.

Wedding/portrait/event photographer booking, package selection, location scouting, recurring shoot scheduling.

Schedule one-on-one or small-group personal training, manage packages, intake fitness goals and limitations.

Membership signup, class schedule, drop-in passes, trial-pass conversion for local gyms and studios.

Class booking, package/credit management, instructor preferences, waitlist handling for yoga and Pilates studios.

Read sensor telemetry and actuate devices over MQTT, Matter, Zigbee, HomeKit, AWS IoT.

Compose multi-device routines — lights, locks, climate, media — across Home Assistant, SmartThings, Alexa.

Intake for web/app development projects, scope and tech-stack discovery, schedule discovery calls.

Freelance/agency designer intake for brand/print/web design, scope a project, surface portfolio, schedule kickoff.

SMB marketing services intake: SEO, social, paid ads, content; scope and pricing for monthly retainers.

Curate, describe, recommend, and contextualize visual artworks or media collections.

Generate, edit, or transform images via diffusion or generative models, including style transfer and inpainting.

Design A/B tests, power analysis, hypothesis framing; review experiment quality and statistical methods.

Survey research literature on a topic, synthesise findings, identify open questions and methodological gaps.

Design lesson plans, learning objectives, assessments aligned to standards (CCSS, IB, A-level).

One-on-one tutoring with adaptive difficulty, error analysis, and progress tracking across subjects.

Surface a corner grocery / neighborhood market's current inventory, build pickup orders, answer stock questions, weekly specials.

Build catering menus for parties, weddings, corporate events; produce quotes, manage dietary restrictions and headcounts.

Take pizza and Italian-cuisine orders, customize toppings, calculate prices, schedule pickup or delivery, and answer menu questions for pizz…

Handle coffee, pastry, and light-meal orders for cafés and bakeries: customizations, loyalty rewards, and pre-orders for pickup.

Accept takeout/delivery orders, dispatch couriers, track ETAs, handle special requests across restaurants and ghost-kitchens.

Identify the right part or tool for a DIY problem, check stock at the local hardware store, give basic how-to guidance.

Eyewear selection, frame style guidance, lens options, prescription handling, and appointment booking for optical shops.

Surface a small boutique's clothing/lifestyle catalog, size and fit guidance, in-store availability, hold-for-pickup.

Take bouquet/arrangement orders for occasions, delivery scheduling, sympathy/wedding/birthday options, same-day cutoff awareness.

Help shoppers find books at a specific independent bookstore, check stock, place holds, surface staff picks and events.

Book overnight boarding or doggy daycare, surface availability and special-needs accommodations.

On-demand and scheduled dog walking, recurring walks, key-handoff coordination, photo updates.

Schedule pet grooming by breed/size/coat type, pre-collect vaccination records, manage recurring grooming schedules.