Poe
Poe brings together model-provider bots and user-created bots for text and multimedia outputs.
Uses, limits and trial →Each profile documents features, limitations and a trial to run with your own test data.

212 · Profiles found
General purpose, multimodal and answer comparison assistants.

Poe brings together model-provider bots and user-created bots for text and multimedia outputs.
Uses, limits and trial →Lumo is Proton’s assistant for conversation, file-based work and web search. Proton explains privacy mechanisms in a separate security model.
Uses, limits and trial →TypingMind is a chat interface that connects models using your API keys and organises conversations, prompts, agents and integrations.
Uses, limits and trial →Character.AI lets users chat with fictional characters and create their own by setting a role, greeting, voice or visibility.
Uses, limits and trial →Manus describes a workspace that carries out multi-step tasks and prepares outputs such as slides. Product updates document connectors and Manus 2.0.
Uses, limits and trial →Duck.ai offers conversations with models from several providers. DuckDuckGo says conversations are not used to train those models. This is a remote service: privacy and offline operation are separate questions.
Uses, limits and trial →ChatGPT can work with documents, spreadsheets and presentations uploaded to a conversation. It can help compare, summarize or analyze them. Supported formats, limits and capabilities depend on the plan and account settings.
Uses, limits and trial →Claude offers Projects to collect knowledge and instructions around a body of work. Documents can supply conversation context. Adding context helps guide an answer; it does not automatically establish every statement as a verified fact.
Uses, limits and trial →Gemini Apps can receive files and answer questions about their content. Google’s help distinguishes device uploads from Drive access, which depends on connections and, for some business accounts, administrator settings.
Uses, limits and trial →Mistral Vibe combines conversations, work tasks and coding features. Its documentation describes a shared web and mobile interface for Chat and Work, with a fast or think toggle.
Uses, limits and trial →DeepSeek offers a conversational assistant and an API platform. An online service, API and downloadable model weights are distinct ways of using AI.
Uses, limits and trial →Prelulu offers text, voice and live avatar video conversations with fictional AI characters. Users can create a character and roleplay a scene; the video shows a responding avatar rather than generating an entire scene on demand.
Uses, limits and trial →Grok is a conversational assistant for text, voice, search and visual content. The official FAQ identifies its provider as SpaceXAI LLC, named SpaceXSI.
Uses, limits and trial →Microsoft presents Copilot as a workspace for asking about context and producing outputs with its applications. Distinguish individual and professional workflows.
Uses, limits and trial →The Kimi interface offers conversations, projects and requests to an agent. This profile documents that entry point without assigning performance to a model.
Uses, limits and trial →Genspark presents an AI workspace with writing, document and visual-content tools. Evaluate its workflows separately.
Uses, limits and trial →Venice presents access to several models for text and media, plus an API. Its site distinguishes privacy workflows.
Uses, limits and trial →HuggingChat is Hugging Face’s conversational interface for chatting with AI models. The selected model and context are part of the trial.
Uses, limits and trial →Meta AI is the general-purpose assistant at meta.ai and in the Meta AI app. Meta documents research and slide creation workflows as well as planning with connected services.
Uses, limits and trial →Engines with sources, synthesis, scientific search or web exploration.

Litmaps supports literature discovery through paper maps, visualisation and publication monitoring.
Uses, limits and trial →ResearchRabbit recommends publications and lets readers explore paper relationships from a collection.
Uses, limits and trial →WolframAlpha computes answers using built-in data and methods rather than merely listing web pages.
Uses, limits and trial →Exa provides a web search API returning URLs and optionally associated content or highlights.
Uses, limits and trial →Perplexity combines web search with conversational answers and links to sources. Its different search modes do not replace reading the cited documents.
Uses, limits and trial →SciSpace combines paper search, literature review and PDF question answering. It offers tables for comparing papers.
Uses, limits and trial →R Discovery helps users follow scholarly literature by topic and discover new papers, including through its mobile app.
Uses, limits and trial →Kagi is a search engine that lets users adjust website weighting or block sites from results. These settings can help repeated research but also affect the sources you see.
Uses, limits and trial →Brave Search combines web results and AI answers and can be accessed in your preferred browser. Brave describes its own index and answers with sources. Conventional results and summaries play different roles in verification.
Uses, limits and trial →Consensus focuses on scientific literature. It retrieves publications before producing cited syntheses. Its value is connecting questions to studies; methodological reading of the papers remains essential.
Uses, limits and trial →Elicit provides scientific search and information-extraction workflows for papers. It can help organize results into tables and cited reports. Extraction speeds up reading; it does not guarantee correct interpretation of every study.
Uses, limits and trial →AI Mode is a Google Search experience with generated answers, follow-up questions and web links. Availability depends on workflow and Google’s conditions.
Uses, limits and trial →Scite offers literature search, an assistant and Smart Citations for examining how publications are discussed, supported or challenged.
Uses, limits and trial →Vane is presented in its repository as an AI-assisted answering engine. Model and search configuration are part of deployment.
Uses, limits and trial →Semantic Scholar is a scholarly search engine from Ai2. It provides paper records, citation links and short automated summaries for some papers.
Uses, limits and trial →Connected Papers builds a graph of related publications from a starting paper. Its documentation distinguishes the main graph, prior works and derivative works.
Uses, limits and trial →Undermind searches scientific literature from a natural-language question, ranks references and produces a synthesis. It can analyse articles whose PDFs are available or added to the library.
Uses, limits and trial →Dimensions connects publications and datasets with other research records. The free edition is intended for personal, non-commercial use and supports searching publications and datasets.
Uses, limits and trial →Images, posters, brand visuals, editing and assets.

Black Forest Labs documents FLUX models and their API for generating and editing visual media.
Uses, limits and trial →Meshy offers text- or image-based 3D model generation and texturing functions.
Uses, limits and trial →remove.bg automates image background removal to isolate a subject.
Uses, limits and trial →Clipdrop combines editing tools including object removal, background replacement and image upscaling.
Uses, limits and trial →Leonardo offers image generation and editing, custom models and motion or upscaling features.
Uses, limits and trial →
Topaz Photo offers denoising, sharpening, upscaling and other photographic corrections in a Mac and Windows application.
Uses, limits and trial →Let’s Enhance provides online image upscaling, enhancement and restoration with several processing modes.
Uses, limits and trial →Image Describer generates descriptions, draft alt text and image prompts, and offers text extraction.
Uses, limits and trial →Vignra blurs the background of a photo. Basic processes the image in the browser. AI Refined uses remote processing after confirmation.
Uses, limits and trial →MyEnhancer offers sharpening, denoising, upscaling, restoration and colorisation options.
Uses, limits and trial →Magnific brings together Freepik’s image, video, voice and upscaling AI tools. Creation and editing features sit in the same suite.
Uses, limits and trial →Midjourney generates images from prompts and visual references in a web studio. Variation and upscaling tools can refine generations.
Uses, limits and trial →AnyImg is a browser-based editor for changing photos with written instructions, adding reference images and refining the result in its Studio.
Uses, limits and trial →
Photoroom provides background removal and background creation for product visuals. Removing a background helps present an object consistently; it should preserve the product’s real characteristics.
Uses, limits and trial →ComfyUI is an open-source generative workflow engine organized into nodes. Comfy also provides cloud execution: a visual workflow does not necessarily compute on your own machine.
Uses, limits and trial →Adobe Firefly brings together visual generation and editing functions. The platform lists Adobe and partner models. Model and feature selection are part of the context to preserve for every deliverable.
Uses, limits and trial →Ideogram turns prompts into visuals and provides editing features. Its documentation includes text within images. This focus is useful for testing compositions without guaranteeing spelling or final readability.
Uses, limits and trial →Canva offers prompt-based image generation within its design environment. Generated visuals can be placed in layouts alongside other elements. The workflow’s value includes preparing a complete communication asset rather than an isolated image.
Uses, limits and trial →Recraft offers image generation models and a creative studio. Judge outputs against the final medium, usage rights and editability.
Uses, limits and trial →Krea is a creative suite with tools for images, video and 3D. Evaluate the specific operation rather than treating every tool as one capability.
Uses, limits and trial →Stable Diffusion WebUI offers a Gradio interface for image generation and editing, including text-to-image, image-to-image and region editing.
Uses, limits and trial →Microsoft Designer combines text-to-image creation, design layouts and editing tools including background removal. Microsoft documents personal access with a Microsoft account.
Uses, limits and trial →Video generation, avatars, editing, captions and creative studios.

VEED combines browser video editing, transcription, subtitles and AI-assisted creation.
Uses, limits and trial →D-ID offers talking-avatar videos and conversational visual agents.
Uses, limits and trial →LTX Studio organises video creation through scenes, storyboards, recurring elements and editing.
Uses, limits and trial →Visemix matches filmed mouth movement to an audio track or animates a portrait, depending on the chosen mode.
Uses, limits and trial →Video Upscaler enlarges short clips online by a factor of 2× or 4×, or to a target resolution of HD, 2K or 4K. Standard and Pro use paid credits, with the charge depending on the clip and settings.
Uses, limits and trial →Runway provides text-to-video and image-to-video generation tools and organizes outputs into sessions. For image-to-video, its documentation recommends focusing prompts on the intended motion.
Uses, limits and trial →Pika provides AI assisted video creation and transformation features. Mode, format and options affect results and the ability to revise them.
Uses, limits and trial →Luma offers an AI creation platform including image and video work. The directory’s Dream Machine reference points to this creative environment, whose features evolve.
Uses, limits and trial →Descript provides audio and video editing with transcription based features. Text helps navigation and editing, but listening and viewing remain necessary.
Uses, limits and trial →HeyGen offers avatar video creation and video translation. A credible video preserves the script’s meaning and makes avatar use clear.
Uses, limits and trial →Synthesia is a platform for virtual-presenter videos, including professional communication and training.
Uses, limits and trial →Wondercraft presents a video studio combining generation, avatars, images, voice and sound with an editing and export timeline. Audio and podcast features remain described.
Uses, limits and trial →CapCut presents a video editor with AI-assisted features. Review editing, used media and exports together.
Uses, limits and trial →Google Flow creates video clips from text, visual references, or start and end frames, depending on the mode and model.
Uses, limits and trial →Music, voice, dubbing, audio cleanup and narration.

Murf provides text-to-speech, narration voices and dubbing tools.
Uses, limits and trial →LALAL.AI separates audio elements such as vocals, accompaniment and instruments according to the selected stem type.
Uses, limits and trial →Moises offers musician tools including stem separation, chord finding and speed adjustment.
Uses, limits and trial →AIVA generates musical compositions and documents customisation, audio or MIDI influences and track editing.
Uses, limits and trial →SOUNDRAW generates tracks for creators and artists with workflows aimed at background music or music distribution.
Uses, limits and trial →Cleanvoice automates audio corrections such as filler-word, pause, noise and mouth-sound removal with multitrack editing and timeline export.
Uses, limits and trial →ElevenLabs provides text-to-speech generation. Voice and model selection affect the result: narration should be evaluated in its intended language and publishing context.
Uses, limits and trial →Suno offers music creation from prompts and musical work tools. Available features and track usage conditions depend on plan and creation mode.
Uses, limits and trial →Krisp offers noise reduction and meeting assistance tools. Audio signal processing and meeting summaries serve different needs.
Uses, limits and trial →Adobe Podcast includes Enhance Speech for improving voice recordings. Judge cleanup through listening rather than waveforms alone.
Uses, limits and trial →Speechify presents text reading, dictation and voice-assistance features across devices. Reading and voice production are separate workflows.
Uses, limits and trial →Async, formerly Podcastle, documents multitrack editing in its audio and video editors. Its help centre also describes downloading individual tracks.
Uses, limits and trial →Documents, writing, translation, decks, notes and marketing.

Fireflies transcribes and summarises conversations, with meeting search and action extraction.
Uses, limits and trial →Otter offers meeting transcription and assisted note-taking with summarisation functions.
Uses, limits and trial →Napkin turns text into visuals and diagrams with exports such as SVG, PDF, PNG or presentation files.
Uses, limits and trial →Granola combines notes and meeting context to produce minutes; its desktop workflow uses computer audio.
Uses, limits and trial →Beautiful.ai builds presentations from outlines and Smart Slide layouts that adjust their elements. Its publisher also documents charts and exports.
Uses, limits and trial →QuillBot combines paraphrasing, grammar checking, summarisation and translation with other writing and creation tools.
Uses, limits and trial →Fathom captures meeting conversations and provides searchable transcripts, summaries and action items, with conversation-search features.
Uses, limits and trial →tl;dv offers meeting recording, transcription and summaries for Google Meet, Zoom and Microsoft Teams, alongside business integrations.
Uses, limits and trial →Reclaim schedules meetings, tasks, habits and focus time around events in Google or Outlook calendars.
Uses, limits and trial →Microsoft presents Copilot as an assistant integrated with work content and Microsoft 365 applications including Word, Excel, PowerPoint, Outlook and Teams.
Uses, limits and trial →MarkItDown is a Python file-to-Markdown utility for text processing and LLMs. Available formats depend on installed components.
Uses, limits and trial →Acrobat lets users ask questions about documents, generate summaries and open cited passages. Adobe also documents PDF Spaces.
Uses, limits and trial →Gemini Notebook (formerly NotebookLM) answers using sources added to a notebook. Google documentation distinguishes imported copies and Drive files that can synchronise; access and freshness depend on the source type.
Uses, limits and trial →DeepL offers text and document translation. Available languages, formats, volumes and features depend on the product and plan; examine the workflow matching your actual document.
Uses, limits and trial →Notion AI combines writing, connected-knowledge search, meeting notes and agents acting within a workspace. Custom agents and connectors have their own terms and permissions.
Uses, limits and trial →Grammarly provides writing, correction and rewriting assistance. Suggestion relevance depends on language, audience and text type.
Uses, limits and trial →Gamma creates presentations and visual content. A decision document requires checking figures and references before export.
Uses, limits and trial →Reflect Open organises notes in local Markdown files with backlinks and optional AI through your provider. The page distinguishes Open from Reflect Classic, which remains maintained.
Uses, limits and trial →Jasper presents agents and marketing content workflows. A brief, brand references and output approval remain essential.
Uses, limits and trial →WRITER presents an agent platform for marketing and revenue teams with knowledge, playbooks and work-tool connections.
Uses, limits and trial →Google Workspace presents Gemini features integrated into work applications. Connected environment and permissions determine accessible context.
Uses, limits and trial →Slack presents AI features for finding and summarising work context within its platform. Conversation access is part of scope.
Uses, limits and trial →Atlassian presents Rovo for knowledge search and agent workflows within an organisation’s context.
Uses, limits and trial →Coda’s AI features work within collaborative docs and tables. An AI column can generate output from each new row’s data.
Uses, limits and trial →AI IDEs, APIs, playgrounds, coding agents and model pipelines.

OpenRouter exposes multiple models through a common API and documents provider routing.
Uses, limits and trial →Groq provides a language-model inference API with an interface compatible with OpenAI-style use.
Uses, limits and trial →Zed integrates AI editing and agents with several model access paths.
Uses, limits and trial →OpenCode is an open-source coding agent available through a terminal interface and other clients.
Uses, limits and trial →Pinecone provides retrieval for AI applications: semantic, full-text or hybrid search, metadata filters and namespaces depending on the chosen schema.
Uses, limits and trial →LangSmith collects and inspects AI application traces with quality monitoring, evaluation datasets and user feedback.
Uses, limits and trial →Replicate runs models through APIs with prediction, version, file and deployment management.
Uses, limits and trial →fal documents model APIs, serverless execution and dedicated GPU resources. Inference and dedicated-compute workflows follow different rules.
Uses, limits and trial →Cline is a coding agent available in editors and the terminal; its documentation covers file reading and writing, commands and browser use with action controls.
Uses, limits and trial →Google Antigravity is a development platform with a desktop app and command-line interface. Google documents a separate transition for individual Gemini CLI users.
Uses, limits and trial →Cursor is a development tool with agents for understanding repositories, planning changes, fixing bugs and reviewing code. Its documentation separates these activities and their settings.
Uses, limits and trial →GitHub Copilot provides development assistance through code suggestions, conversation and project work features. Availability varies with environment, plan and organizational settings.
Uses, limits and trial →Aider is a terminal development tool that works with repository files and compatible models. Git integration helps users inspect and retain proposed changes.
Uses, limits and trial →Continue provides code assistance tools in development environments and on the command line. Models and providers are part of its configuration.
Uses, limits and trial →Claude Code is a development agent that can work with repositories and programming tools. Permission controls and diff review are part of its use.
Uses, limits and trial →Warp presents a development platform with Factories, Terminal and Agent CLI. Factories coordinate agents around the software lifecycle; the terminal is a separate product.
Uses, limits and trial →Blackbox presents inference services with dedicated deployments and model routing. The current site goes beyond its initial coding-assistant positioning.
Uses, limits and trial →Bito presents Governor, a layer supplying code context and routing agent requests to models.
Uses, limits and trial →Tabnine documentation describes a coding agent available through CLI or IDE, code chat and administration features. The old homepage redirects to Tricentis.
Uses, limits and trial →The OpenAI API platform integrates models into applications. With a supported model, Structured Outputs constrains content to a specified JSON schema.
Uses, limits and trial →Google AI Studio lets you try Gemini models in a browser, adjust run settings and obtain sample code for the Gemini API.
Uses, limits and trial →Claude Platform provides access to the Messages API. Structured outputs let you request JSON that follows a schema supported by the API.
Uses, limits and trial →The Hugging Face Hub lists model repositories. Model cards can document uses, limits, data and licensing. Inference Providers offers a separate way to run supported models.
Uses, limits and trial →Together AI offers models through a shared API and dedicated deployments. Its catalog helps identify models available for each route.
Uses, limits and trial →Fireworks AI offers a hosted inference API. Its documentation distinguishes arbitrary JSON from responses constrained by a schema.
Uses, limits and trial →Cerebras Inference hosts models behind an API and documents streaming, rate limits and features compatible with the OpenAI client.
Uses, limits and trial →Lovable presents application and website creation from natural-language requests. Review a prototype as software.
Uses, limits and trial →v0 presents application generation, repository connections and publication in the Vercel ecosystem.
Uses, limits and trial →Bolt presents an application and website creation platform with backend and hosting services.
Uses, limits and trial →Replit presents an AI-assisted environment for building applications and websites. Qualify construction, execution and publication separately.
Uses, limits and trial →Kiro presents an agentic engineering environment for software development. State expectations and checks before changing code.
Uses, limits and trial →GitLab presents Duo Agent Platform for agentic software-development workflows within its ecosystem.
Uses, limits and trial →Devin Desktop is the new name for Windsurf. Cognition presents a desktop app that combines code editing, local or cloud agent sessions, and change review.
Uses, limits and trial →Roomote runs development tasks in an isolated environment and can present changes in a pull request. Its official release notes record the move to the Apache 2.0 licence in Roomote 1.22.0 and the removal of the self-hosted user cap.
Uses, limits and trial →Local LLMs, self-hosted interfaces and privacy-first tools.

Jan Desktop documents chat, models, MCP connectors and a local API server; Jan Agent is a separate product.
Uses, limits and trial →GPT4All runs language models on computers and offers LocalDocs for local documents.
Uses, limits and trial →LibreChat is a self-hostable web application for model conversations and configurable endpoints and tools.
Uses, limits and trial →Docker Model Runner manages and serves models through Docker using compatible APIs and registry-distributed formats.
Uses, limits and trial →MLX LM is a Python package for text generation and model fine-tuning on Apple silicon with MLX.
Uses, limits and trial →SGLang is an inference framework for model serving across configurations from one GPU to clusters.
Uses, limits and trial →PrivateGPT is an API layer for AI applications with file ingestion, document retrieval, tools and connectors. The current project connects to a compatible inference server; it does not run models itself.
Uses, limits and trial →KoboldCpp runs GGUF models on a local machine with a KoboldAI Lite interface and an API. The project documents CPU and GPU options.
Uses, limits and trial →llamafile simplifies distributing and running local models through a portable executable. Mozilla.ai documents bundled versions and using external weights.
Uses, limits and trial →TextGen, formerly Text Generation WebUI, provides a local language-model application with an interface, API and different engines depending on installation.
Uses, limits and trial →Docling converts documents into structured representations with PDF analysis, OCR and Markdown or JSON exports. It can run locally.
Uses, limits and trial →OCRmyPDF uses Tesseract to add a text layer to scanned PDFs, making the PDF searchable and allowing text selection.
Uses, limits and trial →Tesseract is an optical character recognition engine for images, available as a command-line tool and a library.
Uses, limits and trial →PaddleOCR offers several recognition and document-analysis pipelines, including options for layout, tables and structured output.
Uses, limits and trial →GROBID parses scholarly publications into TEI XML with elements for metadata, the article body and references.
Uses, limits and trial →The Unstructured Python library partitions files into elements such as titles, paragraphs and tables, which can then be grouped into indexing chunks.
Uses, limits and trial →Magika identifies file types from contents. It is available through local tools, including a command-line interface.
Uses, limits and trial →Ollama runs models locally and also offers cloud features. Its documentation provides a local-only mode using OLLAMA_NO_CLOUD=1. Model selection and configuration determine where processing happens.
Uses, limits and trial →LM Studio runs compatible downloaded models on a machine through a chat interface and local server features. Operations such as model discovery and downloading require network access.
Uses, limits and trial →Open WebUI is a self-hostable chat interface that connects to model engines, including Ollama and compatible APIs. Interface and inference engine are separately configured components.
Uses, limits and trial →LocalAI provides a server for running models on controlled infrastructure and exposing them to applications through an API. Features depend on configured engines and models.
Uses, limits and trial →vLLM is a language model inference and serving library. It is relevant to teams operating an engine and assessing capacity under concurrent requests.
Uses, limits and trial →AnythingLLM provides document-based conversational workspaces with different model providers. The chosen provider determines part of the data flow.
Uses, limits and trial →llama.cpp is a model inference project in C and C++. It is a technical engine; choose the model and its file separately.
Uses, limits and trial →exo groups compatible devices to run a model locally. Its repository documents device discovery, model placement and a management dashboard.
Uses, limits and trial →Automations, workflows, autonomous agents and orchestration.

CrewAI organises agents, crews and flows with tasks, state and control mechanisms.
Uses, limits and trial →Langflow is a Python framework with a visual editor for component-based AI workflows.
Uses, limits and trial →Activepieces offers automation agents and flows through cloud and self-hosted deployment options.
Uses, limits and trial →Haystack combines components, pipelines, document stores and agents for search and RAG applications.
Uses, limits and trial →LlamaIndex provides a framework for ingestion, indexing, querying data and building agents.
Uses, limits and trial →Semantic Kernel is a Microsoft SDK for integrating models and application functions into AI systems.
Uses, limits and trial →Flowise offers a visual environment for assembling AI applications and workflows.
Uses, limits and trial →Browser Use documents browser-driving agents, cloud browser infrastructure and an open-source library. Hosted agents and cloud browsers are separate workflows.
Uses, limits and trial →Agno offers an SDK for agents, teams and workflows, an AgentOS runtime and a management interface. Agents can be served through APIs or MCP.
Uses, limits and trial →Gumloop offers building and managing agents connected to workplace services.
Uses, limits and trial →Lindy presents workplace agents connected to team applications for retrieving context and performing sequences of actions.
Uses, limits and trial →Muse is Meta’s personal agent. It can prepare or carry out certain tasks through a browser and connected apps. Meta documents permissions and confirmations for sensitive actions.
Uses, limits and trial →n8n provides automation workflows and a self-hosted Community edition. Its Sustainable Use License sets conditions: accessible source code is different from unrestricted use.
Uses, limits and trial →Zapier combines application connections with AI automation features. A workflow links events, data and actions; verify behavior in the affected applications.
Uses, limits and trial →Make organizes automations as scenarios connecting modules and applications. AI service connections can participate in workflows while retaining their own access and data processing conditions.
Uses, limits and trial →Dify is an AI application development platform with workflows and knowledge bases. A visual interface still requires flow and permission testing.
Uses, limits and trial →LangGraph is an orchestration framework for stateful applications and agents. Its documentation covers durable execution and human intervention.
Uses, limits and trial →AutoGen is an agent-application framework. Its official repository now announces maintenance mode without new features and directs new projects to Microsoft Agent Framework.
Uses, limits and trial →Microsoft Agent Framework combines agents, workflows and state management. Microsoft presents it as the next generation of AutoGen and Semantic Kernel.
Uses, limits and trial →Pipedream Connect integrates third-party applications through managed authorisation and prebuilt tools. An application or workflow can call these integrations.
Uses, limits and trial →AI platforms, business agents, governance, data and CRM.

IBM watsonx includes products for AI application development, data management and governance.
Uses, limits and trial →Databricks documents development, evaluation and governance capabilities for data-connected agent systems.
Uses, limits and trial →Snowflake Cortex provides AI functions for text and image analysis, including classification, extraction and summarisation.
Uses, limits and trial →Red Hat AI is a portfolio for developing, serving and managing models and AI applications in hybrid environments.
Uses, limits and trial →NVIDIA NIM distributes container-based inference microservices with documented APIs and GPU configurations.
Uses, limits and trial →Glean offers enterprise search, an assistant and agents grounded in connected-source context.
Uses, limits and trial →Dataiku combines data preparation, analytics, machine learning, agents and governance through visual and code-based workflows.
Uses, limits and trial →UiPath combines robotic automation, agents, APIs, document processing and process orchestration with governance features.
Uses, limits and trial →DataRobot offers a platform to develop, deploy and govern predictive, generative and agentic AI solutions with observability.
Uses, limits and trial →SAS Viya combines data, analytics, models and decision deployment with AI lifecycle management and governance features.
Uses, limits and trial →Amazon Bedrock is an AWS service providing foundation models and features for building generative applications. It addresses integration and operation needs beyond a simple chat.
Uses, limits and trial →Microsoft Copilot Studio provides a graphical environment for creating agents and connecting knowledge, tools and workflows. Publishing, connections and governance depend on the configured Microsoft environment.
Uses, limits and trial →Microsoft Foundry provides an environment for developing and operating applications and agents using AI models. It includes evaluation and management features within Azure.
Uses, limits and trial →The consulted Google Cloud page presents Gemini platform for developing, managing and operating agents. The Vertex AI address redirects to this set of resources; qualify components and availability stages in documentation.
Uses, limits and trial →Cohere documents language models and search tools including embeddings and result reranking. Evaluate these components within their intended pipeline.
Uses, limits and trial →Salesforce presents Agentforce for agents connected to business data and actions, including customer-service workflows.
Uses, limits and trial →ServiceNow Now Assist offers generative features tailored to its modules. ServiceNow documents incident summarisation in Now Assist for IT Service Management.
Uses, limits and trial →SAP describes Joule as a conversational assistant with navigational, informational, analytical or transactional capabilities depending on the connected solution.
Uses, limits and trial →Classic engines, useful to cross-check AI results.

Qwant provides web search. This profile focuses on results and source pages; the directory distinguishes this use from AI assisted answers available in the same environment.
Uses, limits and trial →Bing provides a web search engine and spaces for different result types. This profile concerns finding pages and documents, not measuring a conversational assistant.
Uses, limits and trial →Mojeek describes a search engine with its own index. It can provide another source of coverage when checking AI-assisted research.
Uses, limits and trial →Discontinued, migrated, limited or reference-only services.

Needs review
Gemini CLI is Google’s former terminal route for individual users. The free individual, Google AI Pro and Ultra tiers stopped serving Gemini CLI requests on June 18, 2026. Google documents enterprise and paid API-key routes.
Uses, limits and trial →Historical or retired reference
Sora is a retired OpenAI video service. The web and app experiences closed on April 26, 2026. The API closed on September 24, 2026.
Uses, limits and trial →Historical or retired reference
The former GitHub Spark experience on github.com closed in August 2026. GitHub distinguishes previously deployed apps from inference features tied to GitHub Models.
Uses, limits and trial →